RETRACTED ARTICLE: Agenda to adoption: understanding the mechanisms driving fee-free policy development in Sub-Saharan Africa through policy change frameworks

Policy change frameworks are commonly used to understand policy development processes. However, few studies have attempted to apply these frameworks to the recent popular fee-free policy education at the high school level in Sub-Saharan Africa. Investigating fee-free policy development through policy change frameworks can assist both in identifying the genesis of past policies, including who the important actors are, how issues are framed and problematised, and how specific solutions are designed, as well as how to interpret unfolding policies. In this article, we review three prominent policy change frameworks: Baumgartner and Jones' "punctuated equilibrium framework," Sabatier and Jenkins-Smith's "advocacy coalition framework," and Kingdon's "multiple streams framework." After reviewing the frameworks, we apply them to two fee-free policies in Ghana which are Progressive Free Senior High School and Free Senior High School policies to understand the drivers of fee-free policy change. From the socio-political background, three main concepts were derived from these policy change frameworks deducing from the basic assumptions of these theories. They are domestic politics, political and policy entrepreneurs, and socio-economic dynamics. The results show that fee-free policies are largely driven by domestic politics and political and policy entrepreneurs in political executive positions. Factors under socio-economic dynamics are only scope conditions that are not significant to trigger the adoption of a fee-free policy.


Introduction
Programs and policies are subject to change. Change is a common phenomenon in our everyday life. In the public domain, policy change is a recognition of new and innovative ideas or "incremental refinement of earlier policies" (Bennett & Howlett, 1992, p. 275). The nature of public policies subjects them to perpetual flow where elite decision-makers adjust policies according to their perceived interest shaped by electoral competition, socioeconomic status, perceived forces and modern ways of doing things. What makes these characteristics more unpredictable about public policy is that there are no defined principles guiding the scope of a particular policy change and the time of change. In this case, several instances and underlying factors unknown to the ordinary citizen may influence a certain policy change at any point in time.
In the contemporary era, upper secondary education 1 is identified to be an integral part of the quality of a country's human resources (Becker, 1993). In Sub-Saharan Africa (SSA), the rate of upper secondary school participation is low. The average enrolment rate was 35% in 2020 with great disparities between countries (UNESCO Institute for Statistics [UIS], 2021). Meanwhile, the rate of upper secondary school participation is identified to be a key determinant of political participation, a necessary precursor for democratic and institutional quality and has the positive externality of reducing crime in the region (Asante & Bartha, 2022;Shuaibu & Oladayo, 2016;Wantchekon et al., 2015). Notwithstanding, it is argued that the rate of upper secondary education can be increased in SSA through government demand-side interventions by reducing or eliminating the direct cost of schooling called cost elimination policies popularly called fee-free education (Asante, 2022). Cost elimination policies are access-oriented policies designed to offer free education to children and the youth at different levels of education (World Bank & United Nations Children's Fund [UNICEF], 2009, p. 1).
While cost elimination policies at the upper secondary level are being implemented in some countries in SSA (including Kenya, South Africa, Ghana, Malawi, and Uganda among others), Ghana stands out concerning the nature of these policies. Within the political landscape, fee-free policies are changing from one government to the other with different scopes and coverage (Abdul-Rahaman et al., 2018). Progressive Free Senior High School policy (hereinafter: PF) was adopted in 2015 and was replaced by Free Senior High School policy (hereinafter: FF) in 2017. Why is it so in Ghana? What factors can explain the relatively long stability and abrupt changes within the past seven years? Few studies have attempted to answer these questions (Adarkwah, 2022;Mohammed, 2020;Mohammed & Kuyini, 2021), but either focused on only one of the policies or failed to open the "black box" to uncover the underlying mechanisms leading to the changes. The current study intends to make this happen within the academic discourse by using two fee-free policies implemented by two different governments as case studies.
The study focuses on fee-free policy changes at the upper secondary level and aims to uncover the divers or the combination of drivers facilitating these changes. The research is, therefore, embedded within the terrain of policy change theories and education policy literature. The underlying objective is to deliver a weak test of existing public policy change frameworks advanced in developed economies and apply them to SSA using policies adopted in Ghana as case studies. Specifically, we focus to explain the mechanisms driving the agenda-setting and policy formulation and adoption. We aim to extend the conceptual understanding of cost elimination policies from an empirical perspective. In doing so, we first review the policy change framework and education literature to develop key concepts that can be operationalised, and hypotheses formulated within the scope of the study. Afterwards, we gather empirical data and organise them under themes to test our hypotheses. The rest of the paper is structured as follows. In Sect. 2 we discuss the analytical framework from three chosen policy change theories vis-à-vis education policy literature. Section 3 specifies the research question, operationalisation of our derived concepts and formulate our hypotheses. Section 4 provides details about our research design and methodology. The case study results are presented in Sect. 5, and Section 6 is devoted to a discussion and conclusions, and avenues for further research.

Theories of policy change
Public policy change involves structural changes; it has distributional consequences vis-àvis the allocation of resources (Reich, 1995). Why policies change, when and what scope is a tricky question which is not well understood in many spheres of policy domain (Rodrik, 1996). According to United Nations Education Scientific and Cultural Organisation (UNE-SCO) "a policy is a broad statement that sets out the government's main goals and priorities. It is in line with the country's constitution and can be sector-wide (e.g., education sector policy) or specific to a sub-sector or system (e.g., primary education) or a certain issue (e.g., low enrolment rates)" (UNESCO, 2013, p. 7). The low enrolment rate characterising upper secondary education in SSA presents a social problem. This, according to Bhuwanee (2006), is largely influenced by the policy preferences of governments and their development partners. Governments may adopt a deliberate (regulatory) policy to be applied between the cycles to increase access.
However, many policies-even dysfunctional ones-are going through long periods of stability before they change. The underlying question surrounding the above proposition is how can change eventually come at all? Fortunately, policy change frameworks have emerged to explain changes in public policies on how they get unto the agenda table and subsequently get adopted. These frameworks provide a spectrum of explanatory approaches and are the place to look for explanatory factors in explaining policy change [see (Capano, 2009;Capano & Howlett, 2009)].
In this study, we use three policy change frameworks. We first consider a policy subsystem relating to upper secondary education. Policy subsystems can be associated with the punctuated equilibrium framework (PEF) (Baumgartner & Jones, 1991), although it can be associated with other theories as well (Sabatier, 2007). For this reason, we utilise the PET framework in this study. Furthermore, the realm of education involves several interests with both domestic and international coalitions. This is not surprising because of the global vision of Education for All advocated by world leaders (United Nations, 2020). It is, therefore, important to consider the advocacy coalition framework (ACF) (Sabatier & Jenkins-Smith, 1993;Boasiako & Asare, 2015) argued that the education system in countries such as Ghana is characterised by a weak institutional framework and is ambiguous. Policymaking under ambiguous conditions is usually associated with multiple stream frameworks (MSF) (Kingdon, 2014). Apart from the reason provided for each of the selected frameworks, we consider these frameworks to be better choices for inclusion than other frameworks because of their wide use and recognised importance in the field of policy change studies (Sabatier, 2007).

Punctuated equilibrium framework (PEF)
Punctuated equilibrium's basic assumption is that political processes including policies are generally a gradual process that happens at a slower pace. They are stable and incremental processes that usually produce some form of departure from existing systems (Baumgartner & Jones, 1991, 1993Tushman & Romanelli, 1985). PEF explains that policies go through long periods of stability or change incrementally over time because large governmental systems are by nature conservative and favour the status quo. The stability of policies is reinforced by negative feedback which pushes back against any radical change to keep control of the policy. The dominant monopoly in the policy subsystem stops the negative feedback when political institutions (policy venue) and empirical facts, emotive appeals, people's beliefs, values, understanding, and aspiration (policy image) about a certain policy area (policy subsystem) change (Cerna, 2013;Zahariadis, 2007, p. 155). These mechanisms move the issue beyond the expert-dominated monopoly's control. The image can be shaped subtly by purposeful changes in framing or can shift because of a precipitating event like a major crisis or accident. When the image changes, new actors enter the policy-making arena and changes can swiftly occur. External events, political factors, and the role of personalities come along in the conceptualisation of PEF in the form of external forces, policy actors' beliefs, values, aspirations, and their framing and reframing of issues, respectively. PEF is criticised for its inability to explain the actual processes that will lead to how the changes will occur or when they will occur (Cerna, 2013).

Advocacy coalition framework (ACF)
The advocacy coalition framework (Sabatier & Jenkins-Smith, 1993) specifies changes in public policy occurring from sets of core beliefs and ideas about causation and values in public policy. Different beliefs are linked through interactions between external changes or shocks to the political system. The successful advancement of the coalitions' ideas and beliefs can cause change to occur. There are four basic premises of ACF (Sabatier, 2007). First, understanding policy change requires at least a decade of perspective. Second, policy subsystems are made up of actors from a variety of organisations concerned with an issue. Third, the policy subsystem includes actors from all levels of government. Fourth, policies are like belief systems, and policymakers hold strong beliefs and have a strong motivation to turn these beliefs into policy. One other fundamental truth of the ACF concerns the relative impact of actions internal to the policy subsystem (policy-oriented learning) versus outside events (external perturbations). A common feature that can be associated with the conceptualisation of ACF is the influence of external changes and the role of individuals or personalities (their motivation/interest, beliefs, ideas) in shaping policy change. However, determining the beliefs of the actors is not clear due to the inherent interest of individuals (Cerna, 2013).

Multiple stream framework (MSF)
In explaining the process of policy change, Kingdon (2014) distinguishes three streams, namely problem, policy, and politics. The problem stream is the collection of stress on a system and the discourse that accompanies this buildup. Much of the problem stream is defined by how pivotal actors frame issues, meaning that the nature of the problem is contested. In framing the issues, indicators for example figures or dramatic events or crises can be used to assess the existence or otherwise and the magnitude of the problem, if it exists. The policy stream, according to Kingdon, is ideas floating around in the "policy primeval soup" (2014, p. 117). It is populated by the knowledge/perspectives around some policy problems. In this stream, policy proposals are created and recreated by people with technical knowledge about the issue. These experts range from academics, researchers in think tank institutions, lawmakers, bureaucrats and other domestic or international agencies who share a common aspiration and concern on a policy area. Finally, the political stream is defined by election results, the makeup of legislative bodies, and the prevailing national mood or public opinion around an issue.
Although these streams work largely independently of each other with their distinctive dynamics and rules, they come together at certain critical points, called coupling. The coupling presents a fleeting opportunity "for advocates of proposals to push their pet solutions, or to push attention to their special problems" (Kingdon, 2014, p. 203). This creates the policy window and policy entrepreneurs emerge for change to successfully happen. Policy entrepreneurs are individuals interested and willing to invest "time, energy, reputation, money-to promote a position in return for anticipated future gain in the form of material, purposive and solidary benefits" when the policy window opens (Kingdon, 2014, p. 179). They seize the opportunity immediately to initiate actions. MSF has an inherent quality of explicitly simplifying the causes of policy change-the existence of independent streams, the window, and entrepreneurs-which offers a practical understanding of reality with many variables (Capano, 2009).

Comparing and contrasting the three frameworks
A quick overview of the three frameworks demonstrates that there are clear points of overlap and clear differences. For example, these frameworks all share a goal of organising investigation grounded in theory and aim to offer explanations for policy phenomena (Schlager, 2007). As Schlager (2007) notes, "frameworks bound inquiry and direct the attention of the analyst to critical features of the social and physical landscape" (p. 293). In addition, these policy frameworks help to clarify general variables at play in policy, both in terms of agenda-setting to policy adoption. Third, each of the frameworks emphasises collective action and the importance of certain individuals in leading efforts (e.g., policy entrepreneurs or coalitions).
However, important differences also exist between the three theories. For example, each of the three theories differs in terms of scope/scale (Schlager, 2007). The PEF illuminates the "most coarsest scale" …and "accounts for system-level patterns of decisions or policy adoptions surrounding a policy subsystem" (Schlager, 2007, p. 297). Few variables are identified in the PEF-only interest mobilisations, policy image, and venues (Baumgartner & Jones, 1993) where framing and reframing of issues play a key role in changing the policy image (Baumgartner, 2013;Princen, 2013). ACF explains neither patterns of decisions nor particular policy adoptions. Instead, it endeavours to explain changes in a policy subsystem over long periods (Schlager, 2007). The MSF, by contrast, focuses on why and how specific policy adoptions take place rather than patterns of adoptions (Schlager, 2007). Because the MSF is concerned with specificity, it identifies and deals with more factors and the complexity that comes with "idiosyncratic aspects of policymaking processes" (Schlager, 2007, p. 298). The window of opportunity (labelled as a critical juncture or conjuncture policy window) is identified in MSF which is akin to sudden or major events in PET. The role of beliefs of policy actors in shaping policy ideas is a key concept for both the advocacy coalition framework (ACF) and the punctuated equilibrium theory (PET). Again, both frameworks take into account the theoretical relevance of discursive factors in policy change. Additionally, the ACF approach claims that there is a tendency for policy actors (individuals) to exaggerate both the power and maliciousness of their policy opponents-this is referred to as the devil shift (Sabatier et al., 1987).
Our aim in this study is not to project one framework over the other. However, the goal is to utilise the merits of each of the frameworks in a complementary theory approach. We use the basic ideas in the frameworks identified to cause policy change and deduce the probable factors that cause fee-free policy changes in education (Blatter & Haverland, 2012). To achieve this, we first harmonise the frameworks according to agenda-setting and policy formulation and adoption cycles of the policy process to conceptualise our key independent or explanatory variables.

Harmonising the framework to illuminate key variables
Agenda-setting is about the recognition of a problem on the part of the government (Kingdon, 2014). In the words of Cobb et al., (1976, p. 126), agenda building is the "process by which demands of various groups in the population are translated into items vying for the serious attention of public officials". Agenda setting in the policy cycle can be systemic (unofficial agenda) and institutional (formal agenda). The systemic agenda "consists of all issues that are commonly perceived by members of the political community as meriting public attention and as involving matters within the legitimate jurisdiction of existing governmental authority" (Cobb & Elder, 1983). An example here can be a general understanding that all school-going age populations need to have the opportunity to be enrolled in school. The formal or institutional agenda, on the other hand, consists of only a limited number of issues or problems to which attention is devoted by policy elites (Kingdon, 2014). For example, highlighting the issue of providing fee-free education to all schoolgoing age as a means to increase enrolment.
The framing and reframing of issues, values and beliefs of political agents and coalitions, sudden and focusing events, for example, high demand for high school education, electoral manifestos and outcomes may be identified in the frameworks as means to raise an issue of fee-free education (officially or unofficially) on the agenda table. Manifestos, for example, set national agendas because they play an important role in elections (Caplan, 2007) in which political parties prepare and outline their policies and program proposals. In these factors, the framing of issues, the activities of advocacy coalition, focusing events and the shift in government position can be generally conceptualised as domestic politics. Domestic politics has been a dominant factor in setting policy agenda, making institutions and policy actors more responsive to public issues, particularly fee-free policies in Ghana due to the popularity of these policies (Adarkwah, 2022;Asekere, 2021;Avenstrup et al., 2004;Correa et al., 2020;Frempong, 2020;Grindle & Thomas, 1989;Koop et al., 2022;Mohammed & Kuyini, 2021;Zohlnhöfer, 2009).
Individuals within the political space and coalitions such as candidates seeking political office or education advocates, respectively have interests, beliefs, and ideas that they seek to highlight and put onto the national agenda. These factors can be broadly conceptualised as political and policy entrepreneurs. The role of individuals' interests, beliefs, and ideas in setting national agendas is widely discussed in the literature (Béland, 2016;Béland & Cox, 2016;Brummer, 2016;Swinkels, 2020). Socio-economic dynamics can be used to represent events and problems that arise within the educational system (for example increase in the demand for secondary education) that need the attention of the political community and policy elites and merit public attention. The context in which a policy change is expected to occur is an important matter of consideration in policy studies (Grindle & Thomas, 1989. Policy formulation and adoption involve specific kinds of political institutions capable to develop and refine policy options for the government (Cerna, 2013). The makeup of these institutions "have some minimal level of knowledge of the subject area, allowing them to comment, at least hypothetically, on the feasibility of options put forward to resolve policy problems" (Mintrom & Williams, 2013, p. 19). A reflection on policy formulation and adoption in the policy change frameworks shows factors such as new governments or elites in political institutions, coalitions to influence government decisions, and the political will (window) to adopt a policy to respond to a problem. These factors can be surmised under domestic politics. Like agenda-setting, domestic politics is important in policy formulation and adoption (Koop et al., 2022;Little, 2011).
Again, the interests, beliefs and ideas of main decision-makers, the interplay of individual agents, representatives of coalitions, and the role of policy entrepreneurs are necessary. These variables at the policy formulation and adoption stage within the policy frameworks can also be conceptualised as political and policy entrepreneurs. Individuals including elected and non-elected officials have always played a significant role in the interaction between politics and administration on what needs to be adopted and at what time in the public domain (Young et al., 2020). In addition to this, the literature concerning political patronage and the exercise of personalism in access to public good identifies education as one of the distributive sites individuals use to achieve this purpose (Raby, 1985). Finally, a key factor to consider in policy formulation and adoption is the ability of the macroeconomic condition to support a decided policy option. Also, neoliberalism makes policy options a necessity to be aligned with a global vision, for example, secondary education for all promulgated under SDGs (United Nations, 2020). We conceptualised these factors as the socio-economic dynamics. In summary, domestic politics, political and policy entrepreneurs and socio-economic dynamics are, therefore, proposed independent variables that facilitate or drive fee-free policy change.
We summarise the factors identified according to agenda-setting and policy formulation and adoption in Table 1 and group them under our three main concepts. The pieces of evidence that are gathered in the empirical section are applied to the deduced sets of specific propositions from the abstract theories to identify and explain the mechanism(s) facilitating fee-free policy changes (Blatter & Haverland, 2012, pp. 144-145). To avoid overlaps in the deductive approach, one element appears under one framework although that specific factor may also be at play in another framework.

Research question and hypotheses
The research seeks to answer the following research question: What combination of independent factors facilitates the agenda-setting and formulation and adoption of fee-free policy in SSA? The dependent variable of the study is the outcome of fee-free policy changes at the upper secondary school level in Ghana in 2015 and 2017. The independent variables are the factors identified in the policy change frameworks (according to the focus of this study) to cause policy change. As discussed above, the factors have been grouped under three main concepts. They include domestic politics, political and policy entrepreneurs, and socio-economic dynamics. We operationalise the concepts as follows: Domestic politics Our working definition for domestic politics is the strong ownership through political institutions and coalitions, and the perceived political capital from feefree policies due to the popularity of fee-free policies.
Political and policy entrepreneurs This is defined as the strong long-held belief system, interest and ideas of political and policy actors in fee-free policy.
Socio-economic dynamics The external influence and societal capacity to respond to the identified low rate of enrolment within the secondary school educational subsystem and the macro-economic capability to address the problem.
Accordingly, we specify the following hypotheses: H1 Domestic politics facilitate the agenda-setting and formulation and adoption of fee-free policies.
H2 Political and policy entrepreneurs' strong beliefs, interests and ideas about fee-free policies facilitate the agenda-setting and formulation and adoption of fee-free policies.
H3 Socio-economic dynamics facilitate the agenda-setting and formulation and adoption of fee-free policies.

Case selection
Case studies are considered to be a powerful method for locating causal mechanisms and explaining outcomes (Gerring, 2007). Accordingly, the research is designed as an embedded case study purporting within-case analysis. The unit of analysis is a policy field that is cost elimination policies at the upper secondary school level. Two diverse cases are selected within one country (Ghana) for theoretical and practical reasons. Different categories of cost elimination policies can be identified among countries in SSA in terms of scope and coverage. On scope, one, the policy may eliminate all direct costs of schooling, other school supplies, and ancillary costs (Asante, 2022). For example, Free Senior High School Policy in Ghana (Abdul-Rahaman et al., 2018). Two, the policy may be limited to some direct costs only such as tuition fees and registration fees commonly described as the "big bang" approach (Morgan et al., 2014). For example, Free Day Secondary School in Kenya (James et al., 2016). On coverage, one, the policy may have a national geographical character. And two, the policy may be through targeting or phasing in approach by region or grade/age level or limited to some selected deprived geographical areas, vulnerable students, or for a specific gender (Asante, 2022). Examples include No-Fee Policy in South Africa and Girls' Scholarship Programme in The Gambia (Blimpo et al., 2019;Branson & Lam, 2017). Two cost elimination policies have been implemented at the upper secondary school level in Ghana within the last seven years. They are Progressive Free Senior High School policy (PF) and Free Senior High School policy (FF). The two cases of Ghana stand out with regard to the scope and coverage.
PF was adopted in September 2015. It was a form of partial funding for the reduction of the cost of senior high education for students. Specifically, students were exempted from paying the following costs of education; examination fees, entertainment fees, library, Students Representative Council (SRC) dues, sports fees, culture fees, science development and mathematics quiz fees, Information and Communication Technology (ICT) fees and co-curricular fees for day students in public senior high schools (Ministry of Finance, 2015). In summary, under the PF, the government abolished fees for day SHS students in some selected schools and subsidised boarding fees, gave scholarship schemes to some of the students, and built Community Day Schools to absolve more students into the SHS. Regardless, full tuition and admission fees or registration fees were not absolutely absorbed by the government. This means the policy was limited in terms of scope and coverage, hence a partial fee-free policy. FF replaced PF in September 2017 when a new government was elected to office in January 2017. The policy absorbs all direct costs of schooling and other indirect costs such as feeding and boarding fees throughout upper secondary school institutions in the country to provide holistic free education to all students at senior high school (Abdul-Rahaman et al., 2018).
The diverse characteristics of the two policies afford the opportunity to do intensive two cases set within the same country. In addition, since the two policies are implemented by two different governments with different leading policymakers or decision-makers, the research is less biased towards motivations of a certain ideological belief or individuals' political affiliations within governance. This strategy of case selection can also be justified in the sense that issues in one country in a policy field may vary (Freeman, 1985).

Data
The analytical work of this paper is based on primary data (elite interviews) and secondary data. According to George and Bennett (2005), interviews serve as a major tool for gathering basic information about a case under study. Some information may simply not be available. Furthermore, some historical gaps can be filled with interviews. In the interviews with key policy actors, purposive sampling technique was applied to select participants who are knowledgeable and had more information pertaining to the policies' processes. This non-probability sampling technique is selected following Tansey's (2007) argument that in elite interviews studies, interviewees must have been significant players, have a good memory of the event and are willing to disclose their knowledge of events. Eighteen personal (face to face) interviews 2 were conducted by the first author with relevant respondents (Table 2) from November 17 to December 15, 2021, in English Language. The semi-structured interviews-with a duration ranging from 35 to 50 min-centred on participants' perceptions and experiences of the two policies' developments with the questions guided by the deduced concepts from the policy change frameworks. The interview guide questions were prepared on A4 sheets for each respondent and the responses were recorded by taking notes. We did a pilot interview before the data collection, which prompted some changes in wording.
We supplemented the interviews with secondary data. The secondary data included published media sources, parliamentary Hansards and other official government documents, political parties' manifestos, official and working papers of international development partners (UNESCO and World Bank) and advocacy coalition policy papers. Documents are often used together with other research methods as a means of triangulation which is used by "researchers [to] make use of multiple and different sources, methods, investigators, and theories to provide corroborating evidence" (Creswell, 1998, p. 202). In this study, we do not attempt to evaluate the effects of the changes in fee-free policies on the educational system or the policies' impact. Fee-free policy is looked at by taking the big picture: drivers of fee-free policy changes within the educational system at the high school level are in focus. A more refined analysis is not carried out.

Data analysis plan
In analysing the data gathered to determine the congruence of the empirical reality with the ideal factors, we employed thematic analysis. This inductive qualitative inquiry involves summarising data on open-ended questions and information gathered from documents into themes relevant to the research ambition (Patton, 2015). Here, the data from the broad perspective are translated into the narrative process, organised under themes, and examine how they support or otherwise the hypothesised elements under consideration. In doing so, we closely read the interview transcripts and transcribed all the interviews. The first author then generated codes from the transcribed interviews. Summaries of each identified theme were noted and examined commonalities and differences in each interview transcript and searched for the overarching story under each explanatory variable. In examining interview sixteen, there was data saturation (Bryant, 2014) indicating no new information was emerging from the interviews.
Validity and reliability of the instruments. The terms validity and reliability are contested when it comes to qualitative research. Validity, especially, has many different denotations and connotations (Creswell, 1998). Even though the terms have different values and content in qualitative research compared to quantitative research, it is still important for the qualitative researcher to prove the credibility and quality of the research. In doing so, we first tested the interview instrument by conducting a pilot interview. Inclusion criteria for the interview participants as well as information from secondary sources were set. The selection of interview respondents was based on institutional affiliation as defined in Table 2. The institution is contacted to introduce the study to them. After they agree to the interview, they nominate a high-profile level person or people within the institution who is willing and has enough knowledge and personal experiences about the two policies to interview. This reduced personal preferences to interview certain individuals. Documents and media sources published in English language that can be traced from reliable sources were the ones included in the study.

Ethics
A letter of authority was obtained from the first author's institutional affiliation to conduct the research. Informed consent was sought from all participants before the interview. To preserve the privacy of the participants, we have removed all names and replaced them with institutional affiliations. In addition, given that some participants hold political appointments and needed to be sure their comments would not adversely affect their carrier, a further extra layer of anonymity was implemented where we synthesised the results to create a combined story of narratives.

Results
The analysis covers two governments' fee-free policy changes (i.e., Mahama 2015 PF and Akufo-Addo 2017 FF). However, it also gives an account of the previous period (especially from 2008) when a fee-free policy at the high school level was first proposed by a political party seeking political power in order to better contextualise the case.

Domestic politics and fee-free policy changes
The 1992 Constitution of Ghana as a legal framework highlights the possibility of providing fee-free policy to all qualified Ghanaians at the high school level. It states in Article 25 (1b): (b) Secondary education in its different forms, including technical and vocational education, shall be made generally available and accessible to all by every appro- priate means, and in particular, by the progressive introduction of free education (Ghana, 1993).
However, the idea of free senior high school was only widely discussed and gained public attention when one of the two leading political parties in Ghana, the new patriotic party (NPP) promised fee-free senior high school during the 2008 general elections (see Citifmonline, 2017 publication). 3 Basic education (primary and lower secondary school) has been made free since 2005 (Akyeampong, 2009). Although the 1992 constitution makes provisions for the consideration of free secondary education, no timeline for the adoption of the policy was explicitly stated (Ghana, 1993). 4 The NPP party raising the issue in their manifesto during the 2008 presidential election made the issue widely discussed and set it onto the agenda table. 5 Once the presidential candidate of NPP made the campaign promise in 2008, the public and all other political parties became highly interested in it. 6 In the NPP's 2008 manifesto, it stated: "…guarantee access to all children of school-going age to free quality education at the basic school level. This will be extended to cover Senior High School education as well [emphasis added]" (New Patriotic Party, 2008, p. 69). However, the party lost the presidential elections after they have governed for two consecutive terms (i.e., 8 years). The other main political party, National Democratic Congress (NDC), in the 2008 presidential election had already launched its manifesto (Ghana News Agency, 2008). In this manifesto, NDC did not promise any form of free education at the high school level. The NDC criticised the idea of fee-free education at the high school level on grounds of its financial feasibility and twisted the whole issue along political lines. 7 Between 2009 and 2012 when NDC was in power after winning the 2008 elections, a fee-free policy was not adopted in the country. This may be explained by the fact that NDC did not promise feefree education during the last elections.
The NDC shifted its position during the 2012 presidential elections. 8 The party raised the issue of fee-free education at the high school but only differed from NPP in terms of the scope of the policy. 9 The party focused its debate on the constitutional provision of "progressive introduction" in which they promised that "the government of the NDC… is equally committed to the progressive introduction of free secondary education under Article 25(1) (b) [emphasis added]" (National Democratic Congress, 2012, p. 20). NPP maintained its free education promise in their 2012 manifesto. These events from both political parties may explain the break from the long stability of fee-paying of senior high school fee funding to the consideration of national fee-free policies. This is because the two major political parties who had always won political power since 1992 (Ghana's fourth republic) have the idea of fee-free education in their manifesto. Since then, the idea of free 3 Interview with a representative of NPP; interview with a representative of a CSO. 4 Interview with a representative of a CSO; interview with a representative of UNESCO. 5 Interview with staff of ministry of education; interview with representatives of CSOs; interview with relevant government official; interview with a representative of NPP; interview with NPP member of parliament; interview with an educationist; interview with representatives of teacher unions. 6 Interview with a staff of ministry of education; interview with a representative of CSO; interview with a representative of UNESCO. 7 Interview with a representative of a CSO; interview with a representative of UNESCO. 8 interview with a representative of a CSO; interview with a representative of NPP; interview with representatives of teacher unions. 9 Interview with a representative of UNESCO; interview with a representative of a CSO. senior high school has dominated political, social and economic talk shows in the country (Kwarteng, 2021). It can be argued that the NDC party/government shifted their position on the idea of a fee-free policy which they have opposed during the last (2008) elections.
The framing and reframing of fee-free policy within the political space due to its incorporation in political parties' manifestoes made the idea highly discussed in the public domain. 10 In addition, coalitions with an interest in education have had advocacy through press conferences and symposiums on the need to improve the level of educational participation since majority of youths at their schooling age are not schooled. 11 But, notwithstanding these advocacies and the fact that the idea of free education is a constitutional requirement, its being on the agenda was facilitated by the political party manifesto promises. 12 When NDC won the 2012 elections, free education policy was not adopted until 2015, i.e., PF. The government envisaged that if they do not introduce some form of a fee-free policy before the next general election i.e., in 2016, they may lose votes. 13 The government introduced the policy one year before the next general elections to demonstrate to the people that they also believe in free education as they promised because many people were enthused about the issue. 14 The NDC government was not a new governing party in 2015 when they introduced PF because they were in office since 2009. When the party promised progressive fee-free education in 2012, it took them three years to adopt the policy. As argued in the literature (Correa et al., 2020;Fernandez & Rogerson, 1995) and confirmed in this study, fee-free policies are highly popular among the youth (students) and the old (parents and guardians) and attracts votes. 15 Consequently, the NDC government deemed it imperative to adopt fee-free education in the run-up to the 2016 elections to fulfil their campaign promise made in 2012. 16 However, with FF, the election of a new government was instrumental in its formulation and adoption. The NPP government was elected in December 2016. It took office on January 7, 2017, and adopted the policy in September 2017. The government envisaged that they have made the fee-free policy their flagship campaign promise since 2008, and they must adopt it immediately after winning political power. 17 The interest of the new government and political institutions headed by a "strong" Minister for Education propelled the policy window in the form of political will to adopt the holistic fee-free policy. 18 In both PF and FF, parliament as another major political institution had no difficulty approving the budgets presented to them by the executive to fund the policy. This is because the political party forming the executive arm of government in these periods has always had the required majority in parliament to approve the expenditures. 19 10 Interview with political party representative; interview with a representative of a CSO. 11 Interview with a representative of CSO; interview with staffs of ministry of education. 12 Interview with staffs of ministry of education. 13 Interview with a representative of a teacher union; interview with representatives of CSOs; interview with a representative of NPP. 14 Interview with a representative of a teacher union; interview with representatives of CSOs; interview with a representative of a political party. 15 Interview with a representative of a teacher union; interview with representatives of CSOs; interview with educationists. 16 Interview with a representative of CSOs; interview with NPP member of parliament. 17 Interview with a representative of CSOs, interview with representatives of teacher unions; interview with a representative of NPP; interview with staff of ministry of education. However, in the formulation and adoption of fee-free policies, coalitions are least consulted. These bodies are only consulted when the policy is already adopted. In the course of its implementation, gaps may be identified where coalitions will raise concerns about these gaps, but their concerns are least considered since the government may have already decided to adopt a certain policy. 20 In summary, domestic politics dominated by political party interest for electoral advantage with strong political institutions were instrumental in the agenda and the formulation and adoption of the two fee-free policies. This means we cannot reject Hypothesis 1.

Political and policy entrepreneurs and fee-free policy changes
Beliefs, interests and ideas of personalities are important for a policy change according to policy change frameworks. The personality of Nana Addo Dankwa Akufo-Addo, the then-presidential candidate of the NPP in the 2008, 2012, and 2016 presidential elections, held a strong belief in the idea of fee-free policy at the high school level. 21 On all campaign platforms during these periods, the idea of a fee-free policy was his main campaign message to the electorates and tagged this promise as "the Sacred Promise" to Ghanaians (Quaicoe, 2012). His consistency in the idea of fee-free policy made personalities in the NDC party/government (notwithstanding the initial enormous idea of the infeasibility of fee-free SHS) start thinking about the idea and subsequently incorporated it in the party's 2012 manifesto. 22 This is not surprising because when a candidate in an election sees that their opponent has an advantage over a certain policy area due to its popularity, they are incentivised to advertise the same issue causing the idea to be high on the agenda table (Dragu & Fan, 2016).
In 2015 when the NDC was in the process of adopting the PF, then president Mahama's ideas and interests were high. During the formulation of the policy, he held meetings at least thrice every week on PF with government officials to get a proper plan for the policy adoption. 23 He was interested in the design of the infrastructure component of the policy and he personally coined the name of those school buildings as the "E-Blocks". 24 The president was seen talking about the introduction of the policy 25 including in his 2014 message on the state of the nation (The Presidency, 2014). The final decision to adopt the policy was through a cabinet decision chaired by the president. 26 The president himself launched the policy to begin implementation (Asiedu-Dartey, 2015). The personality Mr Akuffo-Addo, when in opposition, could not resist the temptation to comment when president Mahama 20 Interview with representatives of CSOs; interview with educationists; interview with teacher union representatives. 21 Interview with a representative of NPP; interview with representatives of CSOs; interview with NPP member of parliament; interview with NPP minister for education; interview with other relevant government official. 22 Interview with a representative of UNESCO; interview with a representative of CSO; interview with a representative of NPP; interview with staff of ministry of education. 23 Interview with former minister for education under NDC government; interview with a representative of NDC. 24 Interview with former minister for education under NDC government; interview with NDC member of parliament. 25 Interview with a representative of UNESCO; interview with representatives of CSOs. 26 Interview with former minister for education under NDC government; interview with NDC member of parliament; interview with a representative of NDC. indicated in 2014 to introduce free education. In a Twitter post he indicated: "If President Mahama is, indeed, introducing free SHS in the 2015-2016 academic year, I would say alleluia" (Nana Akufo-Addo, 2014).
Fast forward to 2017, immediately after his swearing-in into office on January 7, president Akufo-Addo was the first government official to make a public pronouncement on February 12, 2017 on the introduction of free education, barely one month in government (Akufo-Addo, 2017b). Here, he declared his commitment to adopting the policy 27 . During the formulation of the policy, he had some interactions with stakeholders from the CSOs for their ideas after the stakeholder's strong criticisms of the policy in terms of how the policy would be funded. At the meeting, the president indicated his strong intention to kick start the policy no matter the challenges that may be encountered in the implementation. 28 He was not ready to allow anything to delay the adoption of the policy. This position could be explained by his strong belief and the fact that, beyond the constitutional requirement, the idea of a fee-free policy at the high school level was started by him as a campaign promise. The cabinet headed by the president made the final decision to adopt the policy 29 . In summary, the belief, interests, and ideas of political executives have significantly facilitated the agenda and adoption of fee-free policies. We cannot, therefore, reject Hypothesis 2.

Socio-economic dynamics and fee-free policy changes
According to the MSF, policy change occurs when problems, politics, and policies twist to such an extent that existing policy solutions become obsolete in responding to a social issue. As mentioned earlier, the immediate aim of a fee-free policy is to respond to the low rate of enrolment at the high school level. In the mid-2000s, enrolment rate at the basic level of education increased due to other several interventions such as the school feeding programme and an increase in capitation grants (Akyeampong, 2009;Ampratwum & Armah-attoh, 2010;Pajibo & Tamanja, 2017). This event led to an increase in the number of people seeking upper secondary education. 30 An assessment of the ministry of education showed that about 100,000 qualified students who complete basic school could not enrol in high school. 31 Additionally, in 2015 the global focus on primary education for all under the Millennium Development Goals shifted to secondary education for all under the Sustainable Development Goals (SDGs) including upper secondary education (United Nations, 2020). These events put the issue of how to increase enrolment at the upper secondary level high on the agenda. 32 Many reasons are assigned to the inability of people to enrol in high school including poverty, child work, income shocks, inequality and health (Hunt, 2008). However, evidence gathered suggests that the inability of parents to pay for the fees of their children due to 27 Interview with a representative of UNESCO; interview with representatives of CSOs; interview with teacher union representatives. 28 Interview with a representative of a CSO. 29 Interview with NPP minister for education; interview with NPP member of parliament. 30 Interview with a representative of World Bank. 31 Interview with NPP member of parliament; interview with staff of ministry of education; interview with other relevant government official; interview with a representative of CSOs. 32 Interview with a representative of World Bank, Interview with a representative of UNESCO; interview with teacher union representatives. poverty is the major cause of low transition (Amagnya, 2020;Asante & Bartha, 2022). 33 Therefore, the immediate policy solution to the problem is to be able to absorb the direct cost of schooling. Since absorbing the direct cost of fees involves the government's direct intervention (spending) in the economy, the ability to respond is largely dependent on the ability of the macro-economic structure to absorb the cost and later generate benefits from it due to the positive externalities of education on social variables (Asante & Bartha, 2022;Becker, 1993).
Interview participants believe that this social context, that is, the inability of many parents to pay for their wards' fees plays a role in politicians and policy actors thinking of the idea of fee-free policies. 34 However, further evidence suggests that it is not the underlying factor causing policy actors to act. In the 2008 general elections when fee-free policy was put on the agenda by Mr. Akuffo-Addo, the policy was not adopted after the elections because the political party and the personality who made the promise lost the elections, although many qualified students could not transition from lower secondary to upper secondary in these periods. Participants believe that the resource capabilities are not a matter of consideration and provide no sense of urgency or otherwise as a driver to adopt a feefree policy. 35 Furthermore, the macro-economic conditions are less considered during the agenda-setting for fee-free policies and subsequently their adoption. 36 Political actors propose the policy, and it becomes a national agenda without a well-defined source of funding for the policy. 37 Evidence from secondary sources points in the same direction. For example, in 2011 Ghana recorded a high Gross Domestic Product (GDP) per capita growth of 11% and real GDP growth of 14.0% because of the discovery and mass production of oil (Fosu, 2017). In 2012, there was economic growth of 6.7% and 4.8% in 2013. In all these periods, many students were unable to transition from lower secondary to upper secondary due to cost, but a fee-free policy was not adopted. A plausible explanation for this was that the political party constituting the government and the personality in power in these times did not promise any form of fee-free policy during the last elections campaign. In 2015 when the economy was declining (by − 0.17), the government adopted a partial fee-free policy (PF). This policy adoption, according to the interview participants, was to respond to the 2012 campaign promise made by the NDC government since the next elections were approaching (i.e., in 2016). 38 The adoption of a fee-free policy is more of an ideological preference of "who is in charge" and not "what the economy can support". 39 International partners (external factors) are involved in fee-free policies after the government has set out 37 Interview with representatives of CSOs. 38 Interview with a representative of UNECO; interview with representatives of CSOs; interview with teacher union representative; interview with educationist; interview with NPP member of parliament. 39 Interview with a representative of World Bank; interview with representatives of CSOs; interview with staff of ministry of education; interview with teacher union representatives; interview with educationist. 33 Interview with members of parliament; interview with minister and former minister for education; interview with staff of ministry of education; interview with other relevant government official; interview with a representative of World Bank. 34 interview with NPP member of parliament; interview with representatives of CSOs; interview with staff of ministry of education; interview with other relevant government official; Interview with a representative of World Bank. 35 Interview with representatives of CSOs; interview with educationists; interview with teacher union representatives. 36 Interview with representatives of CSOs; interview with educationist; interview with staff of ministry of education; interview with teacher union representatives. the policy, its goals and vision and invites international development partners to support it. In the adoption of the policies, external influences are not a significant determinant. 40 It can therefore be argued that external factors are less significant as a causal mechanism for the adoption of a fee-free policy. In summary, we cannot accept Hypothesis 3 because, although, the low rate of transition within the educational system reveals the need to have a policy to increase access, there is no significant evidence suggesting the socio-economic dynamics facilitates the adoption of fee-free policies.

Discussion and conclusions
In this study, we try to identify and explain the causal mechanism(s) facilitating the agendasetting and subsequently the formulation and adoption of fee-free policies in SSA using two cases from Ghana. In the first step of the study, three potential explanatory factors deduced from policy change frameworks-domestic politics, political and policy entrepreneurs, and socio-economic dynamics-were conceptualised to determine their congruence to the empirical reality. The empirical shreds of evidence gathered were then organised in the second step to establish a causal mechanism. The case research suggests the strong influence of domestic politics and political and policy entrepreneurs acting as political executives as the main drivers of fee-free policies. Individuals in coalition groups have minimal influence when it comes to policy formulation and adoption.
Much evidence from other SSA countries points in the same direction. Like the cases from Ghana, the initiatives to abolish school fees come in fulfilment of campaign pledges in other SSA countries. For example, in The Gambia, President Maa Bio won political power after promising a fee-free senior high school policy which he dubbed as the government's flagship programme. He assumed office in January 2018 and launched the policy in September 2018 (Republic of Sierra Leone, 2018). In Namibia, free education at the secondary level has always been a political topic during elections. Although the constitution only makes provisions for free primary education, the adoption of fee-free education at the secondary level began after the elections of a new government in 2016 (Immanuel, 2017). In Malawi, primary school fees elimination was announced a week between the announcement of the results of the elections and the new school year in the 2002 presidential elections. The new government went into what Fleshman's (2010) study calls "crisis mode," summoning senior officials of the Education and Finance Ministries, donors and others to map out a strategy for implementing the plan. This was deemed important at that critical juncture as the matter was regarded as important during the campaign trail and needed to be acted upon to sustain legitimacy. Other studies on the adoption of universal free primary education in Kenya, Lesotho and Uganda note that the process is triggered by political demand rather than a rational planning process (Avenstrup et al., 2004).
However, to get these policies adopted, there has always been a personality who shows the commitment to ensure that fee-free education gets adopted. In most cases, the political leader, mostly the president leads this new framework and is mostly regarded as bot the political and policy entrepreneur. For example, during the introduction of fee-free education in Sierra Leone, President Maada Bio indicated; "In all these, I must particularly mention the man who inspired me, President Nana Addo Dankwa Akufo-Addo of Ghana. He 40 Interview with a representative of World Bank; Interview with a representative of UNESCO. has shown us the way" (Republic of Sierra Leone, 2018). This follows the launch of FS in Ghana when the president, during the launch of the policy gave a hint of his strong desire and belief to adopt the policy. He stated: Today is a very happy day for the good people of Ghana, for the government, and for me personally. I am here, this morning, to perform a very pleasant task: to launch the commencement of the Free Senior High School policy. When I proposed this policy in 2008, many were those who said Free SHS could not be done… The culmination of that belief, inter alia, resulted in the decisive victory won by the New Patriotic Party and my modest self in the elections of 2016 [emphasis added] (Akufo-Addo, 2017a, para. 1).
Political candidates setting the agenda of fee-free policies within the domestic political sphere during campaigns personalise fee-free policies and make them their flagship policies. After winning political power, they are deeply involved in getting the policy formulated and adopted. This is more viable in a political environment like Ghana where there is a concentration of political power in the executive or the president (Mohammed, 2019). These results are in line with Mohammed's (2020) and Mohammed and Kuyini's (2021) study who argued that elected officials are quick to act on salient domestic issues such as free education presented to electorates during elections.
The socio-economic dynamics can be classified as a scope condition to the adoption of fee-free policies. These are mechanisms that may be inexistent across countries and time and may not necessarily be a factor causing changes within one particular space in time. The evidence suggests that factors within the socio-economic context-the inability to pay for fees hence the low rate of transition from lower secondary to upper secondary, the level of economy and external influence such as the global vision to achieve secondary education for all-are conditions political actors use to draw public attention to set the agenda and to justify the adoption of fee-free policies but are not the underlying mechanisms causing the policy change. In countries where the agenda of fee-free policies dominate domestic politics and where personalities that promise them subsequently win political power is where fee-free policies are adopted. Even the promise of a fee-free policy by one political party or personality leads the other to make such a promise in the same election or the next election. Again, fee-free policies have been adopted both in times of stagnant or declining economic growth and high economic growth in different countries. Therefore, regardless of the level of an economy, it may not facilitate the idea of fee-free policy change to be on the agenda table or be adopted. The results from this study validate the second part of Adarkwah's (2022) findings which argue that the underlying mechanism causing the adoption of FS policy is more of a political strategy to win power rather than responding to a social issue.

Conclusion
Two main conclusions are derived from the study. One, the two main concepts identified to drive fee-free policies-domestic politics and political and policy entrepreneurs-are deduced from the utility of different policy change frameworks. This means it is still useful to utilise the basic assumptions of different theories and compare their congruence with empirical facts to arrive at conclusions. One may argue that the two concepts identified to drive fee-free policies are all highly rooted in MSF; that is, the politics stream and policy entrepreneurship. However, as operationalised in this study, the belief and interests of political institutions and political actors are factors highly relevant in driving fee-free policies. Beliefs and interests are concepts highly associated with PEF and ACF. In this case, it is important to consider the utility of other theories. Theories can be complementary and not necessarily competing. We believe that each framework can incorporate the other in order to better understand the dynamics of policy changes. Those concerned with following the development of fee-free policies in SSA should make use of prominent policy change frameworks as an analytical approach. Although these frameworks are developed in advanced economies, they are useful to better understand the mechanisms facilitating the agenda-setting and formulation and adoption of fee-free policies in developing economies.
Two, fee-free policies in SSA are largely driven by elections and personalities within the political sphere. Although in many countries (including Ghana) there is a constitutional legal basis to have free secondary education, it is not enough to bring the issue to the agenda table. The agenda is set by politicians seeking political power within the domestic political system. Accordingly, motivated by electoral consequences, governments are required to formulate the policy and adopt it when in power due to the salience, popularity, and inherent political capital in fee-free policies.

Limitations and directions for further research
Qualitative researchers need to be aware of and reflect on the inherent biases and assumptions we bring to research and the need to inform the reader about which side the research favours (Becker, 1967). This study focused on the socio-political dimensions of fee-free policies and used the experiences and perspectives of several stakeholders and other secondary sources to draw conclusions. Others from other fields of study, for example, economics may disagree with us. The interview guide construction is a product of the factors deduced from the policy change frameworks due to the hypotheses driven nature of the study and hence are subject to our biases. Different researchers may interpret the framework differently. Notwithstanding these limitations, the study, through both deductive and inductive approaches, provides seminal insights to theorise, explain and understand the causal mechanisms facilitating the adoption of the recent fee-free policies in SSA.
Finally, the strong involvement of domestic politics, ideological preferences, and interests and beliefs of personalities in the adoption of fee-free policies raise the issue of the future of fee-free policies. This is particularly an issue to consider because of the current framework of fee-free policies where there are no strict laws obliging (subsequent) governments to implement the policy. What happens to fee-free policies when there is a political party (domestic politics) and/or a leader who has no interest, no belief or has no ideological preference for the continuation of the policy? What if (the head of) the executive may have an interest to adopt the policy, but the parliament is not prepared to approve the budget simply because the political party forming the executive does not have a clear majority in parliament to approve the expenditure? Although these scenarios and potential dilemmas are extensively studied in advanced democracies (Christiansen & Damgaard, 2008;Russell & Cowley, 2016), there is a need to study this in developing democracies. Interestingly, for example, after the last general elections in Ghana in 2020, the NPP party forming the executive who had a clear majority in parliament in their first term (2017-2020) could not secure this clear majority in their second term. There is a hung parliament. This study could not delve into these issues. Further studies may focus on the future of fee-free policies vis-a-vis changes in political structures.