Does public consultation affect policy formulation? Negotiation strategies between the administration and citizens

ABSTRACT While public consultation is a signature process of democratic policy formulation, many governments manoeuvre to refract citizen’s opinions or conduct it perfunctorily. Using the case of a medium of instruction policy in Hong Kong, this article unveils the strategies that the state and citizens employ to put their opinion through to the final policy text, during a public consultation process. Recent literature has identified the mechanisms through which individual actors or organisations contribute to broad policy agenda-setting or policy programme development. However, yet to be investigated is how they – sometimes with conflicting interests – collectively negotiate a policy with the state via public consultations. This paper investigates this very phenomenon, building on previous work conducted in the public policy field, analysing 51 government-generated documents through both thematic content analysis and critical discourse analysis. The paper uncovers four strategies adopted by administrations (non-commitment, case closure, disengagement for irrelevance, and placation) to evade citizens’ equity-oriented demands and stakeholders’ three counter strategies (mobilising other stakeholders into a coalition, reopening the case pointing out a new problem, and appealing by affirming relevance). The state’s discrete refusals and stakeholders’ conjoint reengagement tactics draw our attention to the complexity and subtlety involved in negotiation via public consultations.


Hong Kong's medium of instruction policy and its actors
This study analyses the negotiation between the state and the diverse actors during the reformulation process of Hong Kong's medium of instruction (MOI) policy via public consultation that occurred between 2005-2010.The original MOI policy was created when the United Kingdom returned Hong Kong (HK) to the People's Republic of China in 1997 upon the expiry of its lease.This momentous handover under the 'One Country, programmes, which used to be the realm of the state (e.g.La Londe, Brewer, & Lubienski, 2015, Olmedo, 2014).Any actors' contribution, however, tends to be confined within the boundary set by the state even in this era of fragmented, heterarchical governance, when the government tries to incorporate demands from diverse actors who may not necessarily share interests.It is this complex negotiation between the public office and the multiparties in formulating the final policy text which has yet to receive due attention (Gunter, Hall & Mills, 2015, Roberts, 2004, Silverman et al. 2019).In particular, the uncoordinated, collective shaping of the policy formulation -in rare occasions mixed with temporal and transient collaboration around an agenda -demands scholarly attention.
This research gap is more obvious with specific reference to the negotiation between the state and stakeholders in formulating the policy text via the public consultation process.In many contexts, public consultation is commonly adopted in policy formulation, whether as a political show or as an effective arena for negotiation, being an essential process of democracy (Roberts, 2004).Public consultation is an interactive, governmentmanaged process where a public institute invites either the entire citizens, or stakeholders, to comment on a proposed policy or issues (OECD, 2022, 36, 37).There are a handful of studies on public consultation, which have appeared far-and-wide in the policy literature investigating its contribution to policy shaping (e.g.Costa, Desmarais, andHird 2019, Xue &Diao, 2021).However, they mostly focused on factors affecting stakeholder participation and power struggles over a broad agenda setting, rather than on the textualised negotiation between the state and other actors, or the use of negotiation strategies employed by either party.
This study addresses this identified research gap by tracing how the state and stakeholders push their own position over the equity-related demands of the MOI policy in HK.The study draws on 51 policy documents generated from 2005 to 2010 between the onset of the public consultation and the final policy announcement and dissemination.Adopting thematic content analysis and critical discourse analysis (CDA) tools (Fairclough 1999, 2003, 2013, Johnson, 2015) (see the Analysis section for further details), our research is seeking to address the following research questions (RQs): (1) What were the equity-related demands put forth by stakeholders concerning the MOI policy?(2) How did the administration respond to the MOI-related equity demands raised by the stakeholders, especially when the demands diverged from the government's narratives?(3) What counter strategies were employed by stakeholders to incorporate their ideas into the final MOI policy text despite the government's initial rejection, and with what degree of success?
In the remainder of the paper, we first discuss previous research on policy formulation and describe the design of our study.We then present the findings concerning the negotiatory strategies used by the administration and the stakeholders, and detail the mechanisms through which the public demands were delivered and the administration accepted, refracted or adjusted the demands, before the demands were finally congealed as a refined policy.We conclude by offering the implications of these findings for policy practice and research.

Policy formulation and the voices of stakeholders
Earlier definitions of policy formulation present policy decision-making as a multi-stage process with an exclusive, top-down and technocratic orientation, and limited to a small set of policy elites with little focus on external influences (e.g.Lasswell, 1971).These obsolete frames present the policy stages to be discrete, not noting the iterative relation between phases (Jann and Wegrich, 2007); assume policy decisions to be rational with the public good in their mind (Birkland 2007); and consider that policies are made within the 'iron triangle' of the congressional committees, the federal bureaucracy and lobbyists, with the exclusion of the ordinary citizens from the decision making (e.g.Adams 1981; see also Durning 1999).
In contrast, contemporary, post-modern approaches to policy research, such as policy enactment theory, sensemaking theory, and policy network studies (e.g.Ball, Maguire, and Braun 2012, Howlett, 2000, Spillane, Reiser and Reimer, 2002), emphasise how diverse political/social actors contribute to the process, while acknowledging the power the state exerts.They underscore the need for citizens to monitor the policy processes, especially the formulation, considering that 'in a modern, complex, plural society . . .[the agenda-setting] is often unscientific and irrational' (Ball 2012, 3).It is not always possible that policymakers to obtain necessary information, and sometimes some policies are made with egoistically self-serving goals.Kingdon's (1973Kingdon's ( , 2003) ) as well as the policy enactment studies (e.g.Ball, Maguire, and Braun 2012), capture this unpredictable nature of agenda setting but also in subsequent decision-making processes.They view the negotiation of an initial policy idea by multiple actors, official or unofficial, as an integral part of policy processes.This line of research also notes the ever-blurring division between the government and the stakeholders (Avelar and Ball 2019, Ferrare & Reynolds, 2016, Gunter, Hall & Mills, 2015); however, in this study they are discussed separately, reflecting the binary created in the consultation-related documents analysed herein.

Policy formulation and public consultancy
Research on stakeholders' contribution to policy formulation through public consultation dates to Arnstein (1969).She reveals how citizens' efforts to make any influence on the policy via public consultation can be foiled (see Roberts, 2004 for a historical review of citizen's contributions via public consultation).Silverman et al. (2019) through a neighbourhood revitalisation case in Buffalo, US, extend Arnstein's work and explore the strategies used in the tug-of-war between the policy planner and the citizens.The administration minimised citizen control by making the public meetings of low profile, 'dividing and conquering' or downright dishonouring of the promises made in those public consultation processes.Though delinked from government strategies, they report two citizens' strategies to influence the policy, i.e. contacting the elected officials to leverage insider information and create an informal patronage channel, or creating and maintaining their presence in LegCo meetings and the media.Scott (2017) notes another stakeholder strategy, that is, to gather feedback while engaging with other actors and use it to repackage their policy solution to make it more palatable to the administration.A few HK-based studies, e.g.Lau (2020), Tollefson (2015) and Morris (1996), also shed light on the negotiation between the state and the stakeholders, and the factors affecting the process.Morris (1996) shows how the government uses the entity created for public participation in education policymaking as a tool to deflect criticism against the government, rather than as a channel to collect and reflect public views on education policies.Tollefson (2015) highlights how the mass media turned the school MOI matter into an issue of national and cultural identity, equity, and fairness, to successfully influence the MOI policy drafting.Lau (2020) reveals how the wider political, social, and economic circumstances interact with education policy decision-making, through a qualitative review of policy documents and empirical studies from the colonial  and postcolonial (1997 onwards) periods.
To sum up, these highly relevant studies set a solid foundation in understanding the government's manoeuvres and the power and agency of the actors who entered the policy discussions.While very insightful, the past research has scarcely examined the subtle negotiation between the administration and the public during public consultation processes, their strategies and their ultimate influence on the final policy text, particularly when they have conflicting views.

Data
This paper draws on all 51 documents generated around the public consultation, which are publicly available on the LegCo and Education Bureau websites (listed in Appendix in chronological order).First, all documents concerning the MOI policy were downloaded from the LegCo website, and then documents that were generated by the implementor of the policy, i.e. the Education Bureau, were also traced from the Education Bureau website.There are four types of documents: 1) the proceedings of the LegCo meetings (five public and two close-door) that were recorded verbatim and made available to the public; 2) 15 publications generated for and from these meetings (e.g.stakeholder reports); 3) 20 documents, including speeches, slides or press releases, that public officers generated to disseminate the policy during seminars or briefing sessions; 4) 9 letters and circulars that the government utilised to provide policy updates to the schools, parents and the public.

Analysis
We conducted a two-step data analysis to identify the equity-related demands and their textual flow through to the final MOI policy.First, to answer RQ 1, we conducted thematic content analysis (Braun andClarke 2006, Miles et al. 2014) on all documents listed in Appendix; any parts presented with words such as '(un)equal', (un)fair", 'discriminating', 'disadvantage' or describes such effect was collated as relevant to equity demands.We also noted each stakeholder's identity and the number and content of the speeches.The identified demands that were not immediately accepted by the state, and the degree that they were accommodated into the final policy text are summarised in Table 1.Second, to seek answers to RQs 2 and 3, on segments identified as relevant to the seven demands, we conducted CDA, drawing on Fairclough's (1999Fairclough's ( , 2003Fairclough's ( , 2013) linguistic analysis tools to reveal how linguistic features are used to frame issues and create impact outside of the document.His intertextual tools, such as genre chain analysis, allow us to follow the interaction between stakeholders and the administration on an issue over time across documents.For example, we traced whether an issue raised by one stakeholder was responded to or referred to within an event and throughout the consultation procedure, and, if so, when, by whom and how, which helps us notice the manoeuvres across events (e.g.omission).To identify the subtle attitude of both parties, we conducted a focused analysis on the speaker/writer's move (action, in Fairclough's term, e.g.request to go beyond a perfunctory action), identity or attitude of the stakeholders who presented or received the demands (identification, e.g.expression of concerns) and positioning of and assumptions about the other interlocuters (representation, e.g.resistant to the proposal) (see Fairclough, 2003 for a detailed explanation of the tools and examples; see also, Johnson, 2015).
To illustrate the genre chain analysis, in the column of Teacher Workload of Table 2, the first row of coding shows that a teachers' union expressed negative views (⊖) on the MOI policy concerning teachers' workload during the first LegCo meetings.The state, represented by the working group, concedes the issue and promises to address it but without specifying how (•).The state (the Education Panel), however, drops this agenda in the two subsequent publications (documents 14, 15).To pressure the government to take follow-up actions, the teachers voice out their negative appraisal of the policy concerning their workload (⊖) in the following two consultation events (documents 16, 22), but more powerfully by linking up with other stakeholders (to be further explained below).The state, represented by the Secretary for Education this time, reacknowledges the issues in his speech for Home-school Organisations and Parent Groups (document 18), though not yet sharing any concrete plans (√), but in the LegCo brief that was published four months later (document 31), concrete, multiple strategies to address the issues were announced (◯).

Findings
This section traces the identified stakeholder concerns or demands regarding educational equity through the three stages of public consultation, i.e., appraisal, dialogue and formulation, and consolidation (RQ1).The administrators' comments on stakeholders' concerns during the LegCo meetings showed that stakeholder views were indeed responded to (RQ 2).However, their demands were not equally reflected in the final policy documents: while the initial opinion-gathering was open to the public and involved a wide range of stakeholders, the outcome selectively reflected these opinions, with some feedback being partially responded to and others being completely omitted across events (RQ 3).
In relation to RQ 1, the thematic content analysis identified the seven interrelated equity-oriented concerns or demands concerning students but also teachers: (i) to ensure equal opportunities for university admission and work, (ii) to reduce teachers' workloads, (iii) to allow for flexibility regarding the 'student ability' criterion in determining who can participate in EMI streams within a school, (iv) to address the labelling effect of diversified MOI arrangements within a school, (v) to respect the right for school choices, (vi) to address the needs of ethnic minority students, and (vii) to mitigate the impact of the school closure policy 3 on CMI schools.
Of the seven equity-related issues identified, the administration and the stakeholders agreed on the demand to provide an equal opportunity for all students to access university admission and the job market.The genre chain analysis revealed that the administration immediately incorporated this demand into all relevant discussions and publications with a positive response and promised to take action.For the remaining six demands, however, the two parties' positions were in tension.The next phase of analysis, concerning RQs 2 and 3, identified the negotiation strategies used by the state and the stakeholders and the outcomes as reflected in the final policy.Depending on their initial assessment of the relevance and legitimacy of the raised concerns at the appraisal stage, the administration took four strategies to diffuse tension and minimise citizen control, i.e. non-commitment, case closure, disengagement and placation.Stakeholders responded to the first three strategies with network mobilisation, reopening the case, and appealing, respectively, while there was no record of resistance to the last one (summarised in Table 2; described below).What CDA revealed is that stakeholders present themselves as respectful of the government when broaching issues, making them more palatable to the government.While this approach may be attributed to the high-power distance in Hong Kong between the government and the citizens as noted in the previous literature (e.g.Morris, 1996), such a phenomenon is observed in other societies including the US (e.g.Scott, 2017).Each of these administration strategies is illustrated below, along with any counteractive agentive acts, if any, through which stakeholders tried to put their demands through to the final policy.

Non-commitment and network mobilisation
The administration's strategy, non-commitment, or acknowledgement of the issues but without elaborating any follow-up actions, was countered with network mobilisation, or gaining support from other stakeholders who do not have direct benefits, for instance, teachers' liaising with parents to solve their workload issue.The concern over teachers' unreasonable workloads raised by the Hong Kong Federation of Education Workers, a teachers' union, during the review of MTT policy [2] was acknowledged and summarised by the Education Commission (now EDB) as a legitimate issue [e.g. 6, p. 43] from the early phase of the consultation.Still, no commitment for any follow-up action was made at the appraisal stage.
In response, teacher representatives persisted with the demand to address the issue throughout the subsequent dialogue phase.While acknowledging the issue once again in a speech by promising to provide supply teachers [13], the administration omitted it from the EDB's proposal for the final MOI policy [14,15].
Undeterred, the teacher representatives continued to demand proper support during the LegCo meetings and, in so doing, mobilised their network with school heads, parents, and LegCo members into a coalition.For instance, in the first public consultation LegCo meeting [2], it was only a teachers' union which discussed the workload.In the subsequent meetings, school leaders, parents, and LegCo members also raised the same issue related to teachers ' workload [e.g. 16, 22].For instance, in the second LegCo meeting, the leader of a parent-teacher association mentioned: Parents welcomed the proposals to fine-tune the MOI policy, but were concerned . . . the operation of EMI and CMI classes within schools would necessitate the preparation of bilingual teaching materials, examination papers, etc.The additional workload arising might adversely affect the quality of teaching.[16, p. 10] In the next meeting, the mobilisation reached the prestigious grants schools' senior management, who voiced their support that 'the council considered it necessary to facilitate teachers in strengthening teaching pedagogies so as to alleviate their workload when switching their MOI in teaching content subjects.' [22, p. 8].The support from diverse other stakeholders led the administration to finally acknowledge language teachers' concern over unfair workloads as legitimate and make a commitment [18], which was reproduced in succeeding genres [e.g. 31].For instance, the Secretary for Education (Head of the EDB) noted the administration's 'need to avoid an unnecessary increase in schools ' and teachers' workload' [18].This commitment was later translated into policy actions.In the final policy, a dedicated section on 'professional support for teachers ' [31, p. 9] outlined the measures to support non-language teachers in preparation for the revised policy.On top of the provision of supply teachers, which was already captured in the original proposal, other measures, such as the production of learning and teaching resources, on-site support, and professional networks, were also listed in this final policy [31].
However, network mobilisation was not consistently effective in eliciting a commitment from the administration.For instance, with the demands over the 'student ability' criterion, the public consultation opened the possibility for change, though without commitment.The initial blueprint for the revised policy suggested by the Education Commission's review in 2005, and subsequently reproduced in different outlets, prescribed that students' ability 'must' be considered in decision-making [14, p. 17].
Stakeholders expressed concerns about that article, noting that students' ability measured at the primary level is not static or innate, but likely to improve over time.Considering the injustice that any premature door-closing would generate, a diverse range of stakeholders (e.g.parents and teachers), urged that English learning opportunities be provided to students of all academic performance levels and in-school streaming allow room for improvement in English proficiency.For instance, parents of more than 10 government-aided schools in 2005 formed a pressure group called Coalition of Education Concerned Parents in response to the government initiatives to change the medium of instruction from English to Chinese and mobilised their network to make their voices heard by the government.The stakeholders also capitalised on existing networks: the Committee on Home-School Cooperation (CHSC) sent a representative to a LegCo meeting demanding that students whose English proficiency was assessed as rudimentary also be given the opportunity to learn (through) English, stating: Attention should be paid to the details instead of the terms of the fine-tuning proposal . . . the students [should not] be allowed to stay in the comfort zone only.Students should be provided with the opportunity to get exposed to English.[22, p.12]The final policy documents [40,41] acknowledged the legitimacy of the claim: Noting the view of individual schools that there may be a possible change in the student intake in future and with the anticipation of improvement in English proficiency of our primary students following our efforts in enhancing the learning and teaching of English at the primary level, we do not rule out the possibility to review the 'student ability' criterion within the six-year period if proven necessary by the education sector.[40, p. 9] The Secretary of Education, however, deferred any immediate action by emphasising the six-year duration of the first cycle of the fine-tuning MOI policy 4 and imposing prerequisites for a policy review (i.e.'if proven necessary').He also distanced himself from the action, referring to it as the responsibility of 'the education sector'.These revealed the administration's unwillingness to commit.These tactics of transferring responsibility and setting a hard-to-meet precondition were also observed in the subsequent documents: The criteria for EMI teaching were adopted in the 2005 Education Commission Report after extensive consultation with stakeholders and academic research and were well accepted by the education sector.Any changes to the criteria should be supported by research and thoroughly discussed and accepted by the key stakeholders in the education sector.[41, p. 5] The then-current bifurcation policy was represented as a practice decided 'after extensive consultation,' evoking authority, and as 'well accepted by the education sector.'In the same document [41], the administration argued that 'the majority of students progressed at a normal pace throughout schooling' (p.8) and that 'no significant differences in public examination results among students in individual EMI schools were noticed' (p.10).This counterargument was a response to the doubts expressed by stakeholders over the appropriateness of assessing students' abilities during admission.Still, this concern found its way into the final policy as a result of network mobilisation, revealing the political-experiential feature of the policy formulation process.

Case closure and reopening the case
In other scenarios, the administration claimed that they have already addressed the issue raised by the stakeholder through the proposed revision, thus closing the case.In such circumstances, the stakeholders reopened the case, by highlighting a new aspect of the very same issue.For the issue of a labelling effect, despite the partial acknowledgement of the stakeholders' demands, the administration claimed that the proposed policy has already addressed the problem, and thus, the case was closed.For instance, the EDB stated [14, pp.5-6]: Since the labelling effect arising from the classification of schools into CMI and EMI is deeprooted in HK society, students who are not studying in the classes under (b) above [referring to the 'by class' arrangement] may feel that they are being labelled.There are views that, by allowing these students to learn a limited number of subjects in English and thereby increase their exposure to English, the labelling effect may be further minimised and the students better motivated in learning.
In the following two public consultation meetings, the stakeholders, however, casted doubt on the effectiveness of the 'by class' arrangement, as it would only reduce the labelling effect among schools to within schools, but 'intensify labelling within schools ' [16, p. 15], thus calling for schools' full autonomy on the matter.
Following these meetings, the Secretary for Education accepted the issue and partially modified the class-by-class division, allowing for other choices such as by subjects or by periods, as noted below by the EDB: During our discussion with stakeholders, some school councils have voiced concerns on how to address the possible labelling effect over the so-called 'by class' arrangement.We have deliberated over these concerns thoroughly and have addressed them in the current proposed framework.The proposed arrangement does not call for simple segregation of classes using either CMI or EMI within a school as envisaged by the EC in 2005.[23, p. 7] The administration yielded and acknowledged the problem of the 'simple segregation of class using either CMI or EMI within a school' previously discussed in 2005.This change was put forward in all the following genres within the administrative setting, including the EDB's policy briefs, speeches, and final policy documents.To illustrate, in the booklet that introduced the revised, 'fine-tuned' MOI policy, the EDB [49, p. 9] stated that: Under the fine-tuned arrangements, schools will NOT be required to adopt a uniform MOI.Nor will they be required to crudely divide their classes into CMI or EMI classes.Schools choosing to adopt their MOI by class must be aware that different classes may need specific MOI arrangements as students' abilities are likely to differ.(emphasis in the original)

Disengagement for irrelevance and appealing
Disengagement for irrelevance refers to the government's withdrawal from a discussion of the demands altogether.While acknowledging the issue as legitimate, the administration attempted to divert attention from the focal policy to other existing policies.For this case, the stakeholders appealed, reaffirming that the case falls within the focal policy.Two concerns, (i) CMI schools' vulnerability to school closure under the school consolidation policy, and (ii) the lack of opportunities for ethnic minority students to attend EMI classes or schools, were not actively engaged by the EDB.Stakeholders' demands were responded to by the EDB on one occasion each at a LegCo meeting where stakeholders were physically present and engaged in a synchronous dialogue, yet the discussion pointed at some other policies.For instance, stakeholders expressed their concern over the adverse effects of the consolidation policy, as it was expected to force around 50 CMI secondary schools to close down in the face of a declining student population [16, p. 15], which drew some attention from LegCo members.Stakeholders demanded immediate action, repeatedly at the following three public consultation meetings: Mr WONG Hak-lim highlighted that HKPTU urged the Administration to abandon the consolidation policy for secondary schools to provide a stable learning environment.(Mr. Wong Hak-lim, HK Professional Teachers' Union) [16, p. 6] To mitigate the concern about under-enrollment arising from the student population drop, she called on the Administration to put the consolidation policy for secondary schools in abeyance until the fine-tuning proposal is put in place.(Ms.Cyd Ho Sau-lan, LegCo member) [22, p. 21] The administration noted it in the document [e.g.11, 41] but did not engage with the issue during those meetings or in other events, ruling it out-of-bounds of the MOI policy.The administration considered that this concern was to be addressed by other policies providing development options for schools with insufficient enrolments, such as merging or collaborating with other schools, undergoing special review, joining the Direct Subsidy Scheme and turning to private operation [42].
The other case that the administration did not engage with is the complaint that the MOI policy disregards linguistic minorities' needs [2,6,22].For instance, a principal and a district councillor noted that the promotion of mother-tongue teaching does not meet the needs of ethnic minority students who learned better in English [22, p. 10, 11].The administration tried to highlight the claim's subjectivity and represented it as unreasonable, and thus, indirectly persuading the readers of the minutes to distance themselves from the demand.The following extract is how the demand was presented [22, p. 11]: [Mr.Mak Ip-sing, District Councilor] was also concerned about education for the non-Chinesespeaking ethnic minority groups who, in his view, should learn better in English.These students should be allocated to EMI classes solely on grounds of ethnicity.(emphasis added) By saying 'in his view' rather than reporting the demand as a general claim, the Secretariat of LegCo presented it as personal and subjective, thereby questioning the view's representativeness.By using the third person possessive, the Secretariat distanced himself from the focal perspective.He also (mis)represented the suggestion as advocating linguistic minorities' participation in EMI courses 'solely on the grounds of ethnicity'.The core of the demands made by the linguistic minority groups was not to impose learning the Chinese language onto ethnic minority children.
In response to the disengagement, in both cases, the stakeholders appealed to reconsider the issues, insisting that they were concerned about the spill-over effect of the MOI policy.Compared to the reopening of the case, stakeholders did not present new reasons but reiterated their claims.By examining the two cases, it becomes evident that the administration utilised the genre chain as a gatekeeping mechanism to exclude those agendas in between public consultation events.For instance, while the administration recorded the call to suspend the consolidation policy in the LegCo meeting logs or reports [2,6], it was not recognised in any of the MOI policy texts, such as circulars produced by the EDB.Instead, the initial claim that the issue was outside of the remit of the policy was reiterated in the final policy.Similar tactics were used by the Under Secretary for Education in response to the linguistic minority's needs, attempting to direct the attention of the public to other policies supporting linguistic minorities: The Administration had put in place a number of measures to facilitate NCS (non-Chinese speaking) students to integrate into the mainstream community.In the coming school year (2009)(2010), the administration had undertaken to increase the number of designated schools 5 to 25.The administration would brief NCS parents and groups in August 2009 on the refined MOI framework, including the support measures for NCS students in learning the Chinese Language.[42, p. 25] The flow of the agenda shows how the administration shaped the process to put forth its own position, despite the public consultation system that seemingly favoured the latter.

Placation
The final tactic used by the administration is the long-standing strategy of placation, where the stakeholders were allowed to advise or plan, and the government accepted the issue as legitimate by engaging with it, but 'retain [ed] for powerholders the right to judge the legitimacy or feasibility of the advice' (Arnstein 1969, 220).Perhaps due to their sheer number, parents garnered the largest response from the administration.Multiple parent associations were invited to different meetings, and each was given an opportunity to voice concerns.They problematised the readiness of the schools to practice the professed EMI policy, e.g.'academic qualification of teachers ' [16, p.7], and quality assurance mechanisms" [p. 13].They also demanded to 'be invited to participate in the discussion and formulation of MOI policies at an early stage' [p.9].Most importantly, they wanted to learn about the eligibility for the EMI stream within the school, in addition to the information about the MOI of each subject.For instance, the representatives of two primary school parent-teacher associations noted in the 2005 consultation meeting that 'The constantly changing methods to fix a criterion of student ability (40%) and the arbitrary figure (85%) of student intake are confusing and frustrating parents who feel helpless ' [2, p. 10].This points to the lack of clarity in the arrangement for EMI vs CMI streams within each school, which would critically affect students' opportunity to learn English, the key ingredient for social mobility.
The administration showed an ambivalent attitude towards the request.During the appraisal and formulation phases, especially when representatives of parent associations were present, the administration expressed their willingness to accommodate their requests.For instance, at a LegCo meeting, the Secretary of Education promised to 'consider the deputations' views and suggestions raised at the meeting' and 'continue to discuss with the stakeholders with a view to further developing the proposals,' and showed their sincerity noting that they had 'reviewed [the policy] in light of . . . the expectation of parents' (emphasis added) [16, p. 13, 14].In a speech to parent groups advocating the revised policy [18], he emphasised that he understood parents' wishes to send their children to EMI schools and expressed his awareness of 'the expectations and the pains' of parents.In another speech to the general public [28], he promised that all the above-mentioned demands made by the parents would be implemented and monitored by the EDB.
What was promised, however, was omitted in the other genres (i.e. in responses given in LegCo meetings and official administrative documents) of the genre chain.It was mostly reduced to the matter of informing the MOI by subjects [14, 23, 26, 31, 42, and 49] or teacher quality in addition [24,40].When all the concerns were mentioned, there was no promise of any action [41].The Secretary even contradicted himself on one occasion, saying that 'parents would not consider . . .EMI classes a school would have, when making school choices ' [22, p.18; emphasis added].In addition, the demand was framed as an outcome of parents' inappropriate attitude, confusion and ignorance or even misinformed obsession, which needs to be addressed through parental education.The administration dismissed the discourse of the right to information, but held parents accountable for disregarding other factors when making school choice: I would like to call upon every one of you to help parents adopt an open and impartial attitude in making their school choices with your professional knowledge.[26, Secretary for Education] Parents will then no longer judge a school merely by its MOI when making their choices of schools.Instead, they will consider more thoroughly whether the schools' ethos, characteristics, facilities and support measures will suit their children's abilities and interest to facilitate their all-round development.[28,Secretary for Education] The audience of the first speech cited above was primary school heads, and the second, the general public.The secretary represented the parental preference for schools with more EMI provisions as a matter of ill-informed judgement.These demanding parents were identified as those whose attitude was not 'open and impartial ' [26] and who had to 'consider more thoroughly' to facilitate their children's 'all-round development' [28].In disseminating the final policy, in principle, transparency of information to safeguard parents' right to information was identified as pivotal, in all types of policy documents, including the EDB's policy briefs, speeches, and booklet.However, the parents' right to information was circumscribed to learning what information was provided and they were denied the opportunity to shape the policy to that effect.Thus, the public consultation occurred in a nominal sense, but not in terms of actual benefits.

Key findings
This paper contributes to research on policy formulation by identifying and examining the mechanisms through which the administration and the public collectively negotiate in the formulation of a policy text.Specifically, it traces the seven equity-related demands over a new MOI policy to bifurcate secondary schools into EMI and CMI schools.In Hong Kong, where English proficiency is a marker of elite social status and indispensable capital to access higher education and prestigious jobs, it is not surprising that the policy proposal instigated controversies over its equity implications.In particular, the paper illustrates the strategies the state and the stakeholders used to put their own views through, when their views were in conflict or tension (see Table 1).
For the six equity-related demands, over which there was a position gap between the two parties, the administration directly or indirectly rejected the demands, while the stakeholders took actions to narrow the gap.We identified four types of strategies that the administration adopted in defeating the demands from stakeholders and three strategies that stakeholders used to persuade the administration.Table 1 identifies the patterns observed in the flow of demands and responses, the relationship among the administrator's initial responses on the equity-related demands around the MOI policy, stakeholders' strategies to put their opinions through, and the degree to which their demands were reflected in the final policy text.
When at initial meetings the administration acknowledged the legitimacy of the demands without commitment, the focal stakeholder group mobilised other stakeholders' support to push for commitment and follow-up actions (the cases of teacher workloads and acknowledging the developmental nature of student ability when streaming students).This finding is aligned with Scott (2017), who found that the 'network of lobbying' enabled stakeholders to influence government agencies' decisions at least temporarily.But this does not necessarily mean that the demands were precisely reflected in the final policy text, as the government may initially exhibit a positive attitude, only to reject them years later (the case of streaming students).Sometimes, the government negated the demands as invalid, claiming that the problem was already addressed by the proposed policy (the case of labelling effect), and thus, closing the case.Stakeholders counteracted by raising a new problem, i.e. the fine-tuning policy will only change the venue of marginalisation from outside to inside of the school, leading to the partial revision of the within-school MOI arrangement.
The administration effectively fended other stakeholder-driven demands with very contrastive strategies.With the two demands, namely, the impact of bifurcation on school closures and the needs of ethnic minority students, the administration acknowledged these issues, but disengaged themselves, framing them as illegitimate for the MOI policy, and pointing at other policies.The stakeholders appealed in both cases, reasserting the potential spill-over effect from the MOI policy on the schools' survival or the minority students' life chances, but the administration remained silent on these issues, and dropped the agenda between events in the policymaking chain.In addition, if they had to respond to the issues due to the issue raisers' physical presence, the administration subjectified the problem.For the final demand to allow parents to make informed school choices regarding the MOI, the government continuously engaged with the issue, given their numbers.In their absence, however, the administration adhered to their original stance using the strategy of placation.The findings have implications for both policy practice and research, which are elaborated below.

Implications for policy practice
Practice-wise, the study points to the extent to which an example of public consultation contributes to a particular outcome.In the focal case, even though the position gap was sometimes not narrowed down through the process, at least the administration acknowledged the issue.Considering the iterative nature of policy (Ball, Maguire, and Braun 2012), the acknowledged but unaddressed issue may later be picked up.Allowing the public to participate in policy formulation also enhances policy capacity within governments by narrowing the understanding gap (Anderson 1996, Dryzek 2002).In a context like HK, where a unicameral system is adopted, such opinion gathering may help the government to further appreciate the public's perceptions of the proposed policy.Such position-narrowing opportunities proved to prevent build-up of civic dissatisfaction with a policy (Arnstein 1969, Roberts, 2004).
Our findings, however, point to the possibility that the gap may sometimes be deliberate, as is often the case in other contexts (e.g.Silverman et al., 2019).The administration brushed off the concern over school choice, blaming the 'improper and misinformed' attitudes of parents; although when the parents were present, the legitimacy of the issue was acknowledged.A political reason may explain the administration's conflicting positions in this case: the government had the dual obligations of demonstrating its commitment to MTT at the handover of HK to the People's Republic of China, while facing the pressure to ensure school leavers' English proficiency (Choi and Kan 2022).If the gap is deliberate, the stakeholders and other citizens need to find effective strategies to (re)engage the government, as the stakeholders of the focal case did, especially if the policy affects the social good such as educational equity.
The findings further highlight the complexity involved in policy negotiations, in addition to its iterative and partial nature, and its political nature.The paper has shown that the complexity may not always arise between the policymakers and the citizens, but among citizens.To review the equity demands and the responses from the diverse groups as captured in this study, equitable educational outcomes may vary among stakeholders.Teachers' increased workload may spell customised care to students and better chances for learning to students; allocating non-Chinese ethnic minority group students to the prestigious EMI schools will decrease local students' chances to enter those schools, and so on.The discursive strategies one group adopted could have served to counter another's.Bundling stakeholders' and administrators' discursive strategies blur the very diverse interests at stake concerning.

Implications for policy research
This paper has identified the detailed negotiatory strategies used by the state and the citizens over an education policy, building on the studies conducted in the field of public policy and using some of its concepts.Thus, it has shown the relevance of the research on public participation conducted in the field of public policy studies (e.g.Arnstein 1969, Roberts, 2004, Silverman et al., 2019) to education policy studies, pointing at the synergy created through interdisciplinary dialogues.We also extend their work by tracing the change across policy events and presenting empirical evidence that public participation via public consultation does influence policy formulation.
Research method-wise, this study shows the potential of combining the thematic content analysis and the CDA, for understanding the policy formulation process using archived data.Through thematic content analysis, researchers identified relevant segments from a large set of data (e.g.reports typically consist of 100-300 pages).On these identified segments, CDA was conducted to reveal the subtle and discreet, linguistic manoeuvres used in the crafted policy texts; the genre chain analysis unveiled how the government avoided conflicts by placating the relatively powerful parents when they were in present or remaining silent on legitimate issues that cannot be addressed.Without combining the methods, the study might not have shed light on the power struggles in the policy formulation processes, and how the public sphere dialogues were made less than 'effective' in the process, in Fairclough's term (e.g. 1999). 6 Finally, related to the first point in the previous section, the study indicates that in theorising the impact of the policy negotiation process, we need to depart from the binary evaluation, either success or failure, but acknowledge partial advancement.Policy narratives subsequent to the public consultation were reasonably congruent with stakeholders' demands on the MOI policy, despite the instruments being not fully aligned with stakeholders' preferred outcomes due to policy histories, a mismatch of policy goals and means, and a lack of a common (professed) understanding of the problem.Even in the cases of disengagement, although the concerns were not positively addressed, the administration acknowledged the concerns openly and listed the policy alternatives to demonstrate its commitment to address the issues by other more appropriate means.

Concluding remarks
In sum, by tracing the seven equity-related demands and analysing the strategies employed by both the administration and stakeholders, this study draws attention to the collective formulation of a policy during public consultation, when multiple stakeholders with sometimes conflicting stakes make uncoordinated, separate demands to the state.It elucidates the government's manoeuvres to reduce citizens' control of the text both immediate and delayed, and citizens' counter strategies which create an incremental but concrete impact on policy text.
The findings highlight the challenges and tensions that arise when policy proposals have unintended but significant equity implications, particularly in a society like Hong Kong where English proficiency is highly valued.By shedding light on the nonlinear, complex equity implications of a policy, it calls for attention of a reform manager to review the enacted policy and take necessary actions to address issues arising.The reform managers will learn of unplanned impact on equity by genuinely listening to stakeholders' views during the public consultations.For citizens, the glimpse into the strategies used by the administration to reject stakeholder demands and the counter-strategies employed by stakeholders to narrow the position gap in this study may help them to best utilise the public consultation, though these strategies are context-bound.
By highlighting the complexities involved, it invites further studies on policy formulation during public consultation.Future studies can explore if there are other repertoires of strategies commonly taken by the two parties, i.e. the administration and the stakeholders, mixed-use of stakeholders' strategies, and whether and how the social contexts affect the repertoires.Also, our data shows that the political orientation of the LegCo members affects their engagement with identified issues, which is not explored due to limited space.Thus, rather than assuming the administration and the public to be homogeneous groups, it is advisable to investigate whether and how individuals' background (e.g.political orientation, position in schools) affects their negotiation strategies.Finally, future studies can incorporate the lack of common (professed) understanding of the equity problem across groups (e.g.local vs minority students; EMI vs. CMI schools), and investigate the tensions among stakeholders themselves, in addition to those between the administrators and the stakeholders.

Notes
1.The move was also perceived to be a means to strengthen students' national identity and patriotic sentiments; however, as a neoliberal global city which prioritises economic development, HK has seen EMI education persist, with English remaining a critical cultural capital for competitiveness and social mobility even after the change in sovereignty (see Choi and Kan 2022 for linguistic ecology in Hong Kong).2. The ongoing Belt and Road Initiative has also opened some changes and opportunities for ethnic minorities.See Choi and Adamson, 2020 for further details.3. Officially termed a consolidation policy, it required schools with insufficient student enrolments to close down or to reduce the number of classes that they operated.As of December 2009, 77 primary schools and 15 secondary schools had ceased operation under the consolidation policy.4. In the LegCo meeting on the first six-year cycle of the fine-tuning MOI (2010/11 to 2015/ 16), it was decided that there will be no revision to the 'student ability' criterion.5. Designated schools were those that were provided with additional resources and focused support to enhance the learning and teaching of NCS students.The system of having schools designated to serve the needs of ethnic minorities was abolished in 2013/14.6.By an effective public sphere dialogue, Fairclough (1999) refers to one that allows access to all, freedom to disagree, space for consensus to be reached and alliances to be formed, and space for discussion that makes a difference.

Funding
The work was supported by the The Education University of Hong Kong.

Notes on contributors
Tae-Hee Choi researches education policy processes and teacher change therein.She also investigates the role of language as an object and medium of policy processes.Choi has led multiple international projects involving countries in Europe, Asia and Oceania, and provided consultancy on policy and teacher development to 13 countries at the invitation of local governments and international organisations (e.g., Asian Productivity Organisation; the British Council

Table 2 .
The flow of disputed stakeholder demands through the policy formulation process.
). Updates on her research activities are available at https://www.researchgate.net/profile/Tae-Hee-ChoiYee Lok Wong is a Research Associate at the Social Inclusion Project, Lee Kuan Yew School of Public Policy.She received her Master in Public Policy from LKYSPP and won the WyWy Gold Medal for Best Student in Public Policy with Best Policy Analysis Exercise.She is interested in the study of social and economic policies to protect people from economic risks and insecurities of life.