Critical success factors for public-private partnership in universal secondary education: Perspectives and policy lessons from Uganda

There is a dearth of phenomenological interpretive studies in public-private partnership (PPP) policy in education service delivery. The limited extant literature on PPPs in education hardly explores insights into how stakeholders understand them, and what they perceive as critical success factors (CSFs) for their implementation in the context of developing countries. The overarching purpose of this study is to explore the stakeholders’ perceptions of PPP policy in universal secondary education (USE) and its CSFs in Uganda. It employed the interpretive paradigm and the participants were purposively selected from government bodies, partnership private schools and local communities. Document review and interviews were used as the data collection methods while the resultant data were analyzed using content and thematic techniques. The findings reveal that most stakeholders’ understandings of the PPP in USE were diverse and context-specific; and that most school-based stakeholders implemented this policy without clearly understanding its origin, goals and guidelines. While most government-based stakeholders perceived the policy as successful, the majority of school-based stakeholders deemed it unsuccessful. The majority of stakeholders perceived regular policy reviews, commitment to partnership roles, sufficient funding, the selection of partners with adequate capacity, effective policy communication, regular policy monitoring and strong enforcement mechanisms as its CSFs. In view of the findings, it can be inferred that unless appropriate policy reforms and best practices informed by these findings are undertaken, the success and sustainability of PPP policy in USE would remain uncertain.


INTRODUCTION
Public-private partnerships (PPPs) have gained considerable attention and popularity internationally as innovative management and financing models for public service delivery and modern infrastructure development for sustainable development (Patrinos et al., 2009;Babatunde et al., 2012;Ginsburg, 2012; *Corresponding author. E-mail: ivantwinok@gmail.com. Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License Moschetti, 2016;Kim, 2017;Badasyan and Riemann, 2020). PPPs enable public and private economic sectors to pool their strengths and expertise within their respective efficiencies and capacities to enhance the delivery of quality public services Boyer and Van Slyke, 2019). Most economies have thus adopted PPPs as neoliberal policy responses and new public management (NPM) practices for promoting greater private sector involvement in public service delivery as government"s role and funding reduce (Ginsburg, 2012;Robertson and Verger, 2012;. Under neoliberalism, PPPs are presumed to enable governments to transfer some operational roles to efficient private sector operators while focusing better on the core public sector responsibilities, such as regulation and supervision (Miraftab, 2004;Olssen and Peters, 2007;Ginsburg, 2012). Thus, PPPs are viewed as viable collaborative arrangements and policy reforms for reducing government"s fiscal and administrative burdens in public infrastructure development and service delivery through increased private sector participation (Cheung et al., 2012;Kim, 2017).
The central philosophy behind PPPs is that all organisations have strengths, but no single organisation has all the strengths required to do everything alone (Rotter and Özbek, 2010). This relates with Reim (2009: 14)"s assertion that "the primary objective of PPP is to deliver a better service than either the public or the private sector could do alone". Thus, PPPs have been adopted as innovative market-led models for improving access to quality public services for the poor and disadvantaged groups cost-effectively through the public and private sectors sharing risks, costs, resources and responsibilities, particularly in capacity and budgetaryconstrained economies (Akyeampong, 2009;Yescombe, 2018). In developing countries, PPPs are perceived as a mechanism for attracting and crowding in private investments and expertise for economic development amidst budgetary challenges and pressure on traditional domestic public sources (Amuche and Kukwi, 2013;Ismail and Haris, 2014;UNESCO, 2015). PPPs reinforce efficiency and synergy through sustainable access to private sources of capital and technology for quick social service delivery and the development of supportive infrastructure (World Bank, 2011;Yescombe, 2018).
PPPs in education have grown globally as a marketoriented approach to improving access to quality education for all (EFA) while ensuring equity and social justice (World Bank, 2011;Verger and Moschetti, 2016). Education PPPs (ePPPs) proponents argue that the rising school enrolments and demand for education as a basic need, human right and public good have rapidly outstripped public sector capacity to manage and provide it sufficiently alone (Rose, 2010;Locatelli, 2107;Termes et al., 2020). Thus, ensuring equitable access to quality public sector education amidst government budgetary constraints requires the adoption of ePPP to reduce the role of the state in its delivery (Srivastava, 2010;Kim, 2017). Likewise, by adopting PPPs in universal education, governments hope to efficiently and effectively achieve EFA goals through increased coverage and equitable access (Robertson and Verger, 2012).
Provision of quality EFA is also viewed as the best equaliser in most societies owing to its distributive power through human resource capacity-building (Amuche and Kukwi, 2013;Ben-Shahar, 2015). Education is thus deemed as indispensable for multidimensional transformation, because investing in the human mind makes all other development objectives possible (World Bank, 2011;Mgaiwa and Poncian, 2016). Education is considered a dynamic tool that enhances national capacity-building and people"s resilience as they strive towards achieving sustainable development (World Bank, 2011). For the aforementioned reasons, ePPPs have won much popularity and growing support as a mechanism for increasing sustainable access to quality EFA at low cost (UNESCO, 2015;Aslam et al., 2017;Moschetti and Verger, 2020).
Nevertheless, the implementation of ePPPs faces challenges (Mahmood, 2013) and their success depends on some contextualised critical success factors (CSFs) (Mannan, 2014). This study provides a PPP literature review from the global and local perspectives, the methods used in this study and the findings on ePPP in the context of Uganda. It concludes with key lessons learnt and policy recommendations.

LITERATURE REVIEW
Globally, PPPs in education gained momentum in the 1980s as a strategic neoliberal option to address the budgetary constraints and other challenges confronting governments" education service delivery systems (Malik, 2010;Kim, 2017). Hence, most sub-Saharan African governments (like Uganda) have incorporated PPPs into the EFA and universal secondary education (USE) programmes to meet the increasing demand for public secondary school education (Patrinos et al., 2009). Besides, owing to persistent failures and equity concerns in education in most developing countries, where the public sector is a key player in financing and providing public services, ePPPs are seen as appropriate avenues for improving public education delivery (Patrinos et al., 2009). This corroborates Akyeampong (2009)"s study finding in Ghana that ePPPs effectively serve the needs of disadvantaged and hard-to-reach children if resources are provided to support them. Relatedly, studies on ePPPs (LaRocque, 2008;Mahmood, 2013;Baum, 2018) indicate that PPPs can reorient education towards improved access, efficiency, competition and quality in its delivery systems and outcomes. PPPs in USE delivery in Uganda emerged, in 2007(Chapman et al., 2010, within such neoliberal educational policy contexts (Patrinos et al., 2009).
While education is viewed globally as a necessary service for enhancing socio-economic development, the resource potential for financing and providing it adequately remains a key challenge, particularly in developing countries (Amuche and Kukwi, 2013;Luthra and Mahajan, 2013). Consequently, some budgetconstrained countries have adopted PPPs as a viable policy option for providing and financing affordable quality EFA. Moreover, the World Bank (2011) opines that in developing countries, ePPPs are justified by the demand for access to affordable schooling and the need to tap into private resources where the state cannot afford EFA. Indeed, most ePPP literature emphasises that ePPPs are a means of increasing equitable access to affordable quality EFA on a sustainable basis (Fennell, 2010;Malik, 2010;UNESCO, 2015;Aslam et al., 2017;Moschetti and Verger, 2020).
According to Education International (2009), ePPP models are categorised based on what kind of education service the state procures from the private sector and how it does this. The common forms of PPPs in education service delivery include: private management of public schools; education vouchers/subsidies and scholarships; contracting out the delivery of education services (purchase of educational services from private schools); school infrastructure initiatives; capacitybuilding initiatives; and education philanthropic initiatives (Latham, 2009;Patrinos et al., 2009;LaRocque, 2011). Though the contracting model seemed common (Termes et al., 2020), another popular model, referred to as buildoperate-transfer (BOT), is emerging. BOT involves large infrastructure projects, where a private sector operator is granted a franchise (concession) to finance, build and operate an educational facility. The government leases the facility for a specified period, after which it is again transferred to the respective state authority (LaRocque, 2011;Mathonsi, 2013;Robertson and Verger, 2012).
The type of PPP in USE that Uganda adopted in 2007 involved outsourcing education service delivery to private secondary schools as partners (MoES, 2007;Patrinos et al., 2009). This form of ePPP is "the process whereby government procures education-related services of a defined quantity and quality at an agreed price from specific providers" (Patrinos et al., 2009:9). In this policy, the government signed a memorandum of understanding (MOU) with selected private secondary schools to deliver USE services to state-sponsored students for a specific time period (Ministry of Education and Sports [MoES], 2013). The government guides policy and provides finance to procure USE services while the contracted private schools deliver them on behalf of the government to a specific number of enrolled USE students, particularly those from poor rural communities with inadequate access to secondary schools (MoES, 2012). This type of ePPP corroborates Srivastava (2011)"s Twinomuhwezi and Herman 135 description of contracting in education as an agreement between a government agency and a private provider to deliver an education service in exchange for regular payment. This study was motivated by the contextual need to gain insights into the less explored phenomenon of PPP policy in USE delivery whose realities had continued to stir up controversies from a wide spectrum of its stakeholders in Uganda (MoES Report on USE National Headcount Exercise, 2014). Besides, while there is a diversity of international and local scholarship on ePPPs (Akyeampong, 2009;Rose, 2010;Srivastava, 2010;Robertson and Verger, 2012;Kasenene, 2013;Mathonsi, 2013;Barungi et al., 2015), it hardly explores how and why stakeholders understand PPP in USE in the way they do, and what they perceive as its CSFs. Yet, for the PPP policy to achieve its objectives and be sustainable, identifying its CSFs is crucial (Mannan, 2014;Rashed et al., 2017). Likewise, because PPP risks and challenges are diverse, identifying context-specific CSFs has been widely recognised as a best practice for the successful implementation of PPPs (Zhang, 2005;Forrer et al., 2010).
Conceptually, CSFs are absolutely necessary for an organisation or project to achieve its objectives (Rockart, 1982;Chan et al., 2010). In this connection, a number of CSF studies (Zhang, 2005;Chan et al., 2010;Abeer et al., 2011;Babatunde et al., 2012;Cheung et al., 2012;Ismail, 2013;Emmanuel, 2016;Fang and He, 2019) have generally pointed out the following as CSFs for the successful implementation of various PPP projects globally: trust; openness; fairness; mutual respect; appropriate risk allocation and sharing; a competitive and transparent procurement process; commitment by the public and private sectors; a favourable legal framework; accountability by partners; government provision of guarantee and political support; stable macroeconomic conditions; and the availability of suitable financial markets.
Though the extant literature on CSFs is abundant, there is limited scholarship on CSFs for ePPPs, particularly on USE delivery in Africa and from the perspective of stakeholders" lived experiences and understandings. Yet such CSFs are relevant both for the effective implementation and management of PPPs in resource-constrained developing countries where PPPs are still new and are performing relatively poorly in public service delivery and for effecting reforms as policy contexts evolve with time (Cheng et al., 2012). Interestingly, empirical studies on CSFs in the extant literature predominantly hinge on non-educational PPP projects in energy, housing, water supply and road infrastructure (UNECE, 2008;Adamu, 2019;Chileshe et al., 2020). Furthermore, a critical review of the different PPP studies reveals that while some CSFs seem to be context-specific and differ across sectors, projects and countries, a few others appear to be universally applicable to PPPs (Cheung et al., 2012;Mannan, 2014;Fang and He, 2019). In this vein, Patrinos et al. (2009) suggest that more research is necessary on the linkages between PPPs and education outcomes owing to different country-specific settings.
Methodologically, phenomenological interpretive studies on the PPP debate are inadequate to enable indepth understanding of this policy phenomenon. Chan et al. (2010) echo this when they point out the need for more in-depth case studies to verify the reliability of the CSFs identified for PPPs in other contexts. Besides, Pakistan Action Aid (2010) reveals that rigorous research on the PPPs is lacking, and that most of the existing evaluation data is generated by the PPP programme owners themselves, who may have conflicting interests. In this regard, some scholars (Patrinos et al., 2009;Ismail, 2013;Jomo et al., 2016) have suggested that there is need for more research and understanding of CSFs for PPPs since the countries in which they are implemented are contextually different.
The foregoing PPP literature indicates that studies on CSFs for PPPs in education based on stakeholders" understandings and perspectives, particularly relating to USE delivery, are missing in this policy debate. Therefore, such contextual, methodological and knowledge deficiencies in ePPP literature justified conducting this study in Uganda, where PPPs are still new and controversial. The following research questions were designed to guide this study: How do stakeholders understand the implementation of PPP policy in USE delivery in Uganda? What do stakeholders perceive as CSFs for implementing PPP in USE in Uganda?

Research design
This study was conceptualized within phenomenology as its philosophical stance, and it employed exploratory qualitative case study approaches through an interpretive paradigm for in-depth understanding of this policy and its CSFs from stakeholders" subjective viewpoints (Crotty, 1998;Yin, 2009;Grix, 2010;Maree, 2012). Moreover, Yuen (2005) argues from the interpretive perspective that understanding is a precondition for correct interpretation and sense-making of the [policy] phenomenon experienced.

Study population and sampling procedures
Wakiso District, in Uganda"s central region was considered as a case study area owing to its dominance in having more partnership schools in rural-urban settings than other districts (MoESTS Statistical Abstract, 2014). The study was informed by Freeman (1984)"s stakeholder theory as its appropriate theoretical lens. The 28 stakeholders for the study, who were purposively selected, included four Ministry of Education (MoES) officials, four district local government workers, two members of Parliament, two school proprietors, four head teachers, four teachers, two parents, two local leaders, two NGO-based educators and two academics with experience and knowledge of PPP policy in USE. Stratified purposive sampling was employed to ensure that an adequate sample of information-rich participants was selected from each specific group of government-based and school-based stakeholders. Snowball sampling also emerged naturally in this study owing to the instances in which some interviewees would refer to other information-rich cases. The sample size was also partly determined by data saturation (Onwuegbuzie and Leech, 2007). Using maximum variation sampling (Christensen et al., 2015), four private secondary schools were purposively selected based on their size and distribution of USE-sponsored students, performance, and location in relation to urban or semi-urban and rural settings.

Data collection and analysis procedures
Qualitative data collection and analysis procedures were employed in this study. Data collection was conducted in 2016 and 2017 through in-depth interviews, document analysis and field notes. Triangulation of data sources and methods, among other trustworthiness measures, was adopted to ensure the quality, validity and credibility of data and research findings (Miles and Huberman, 1994;Maree, 2012;Creswell, 2014). To maintain study participants" confidentiality and anonymity, pseudonyms were employed to represent their responses in the entire study process. The interviews with selected stakeholders were audio-recorded with their voluntary consent. They were transcribed and analysed using content and thematic approaches (Rubin and Rubin, 2011;Saldaña, 2015).The interview data was then corroborated with information drawn from relevant PPP policy documents. The documents reviewed included the MoES (2012) National headcount report on USE, the MoES (2013) PPP policy guidelines for implementing USE in Uganda, the MoESTS (2014) Statistical Abstracts, the Uganda PPP Policy Act (2015), and various journal articles and international publications on PPPs and CSFs. Analysis focused on stakeholders" understanding of PPP policy in USE and their perceptions of its CSFs in Uganda.

Stakeholders' understanding of the policy
This section presents the findings on stakeholders" understanding of the implementation of the PPP policy in USE in Uganda. It focuses on how and why stakeholders made sense of the policy origin, purpose, guidelines and success in USE delivery. It identifies the similarities and differences in the ways stakeholders understand this policy and why. For ease of cross-case comparative analysis of their understandings and perceptions of the policy realities, the stakeholders were mainly categorised into government-based stakeholders (MoES and district local government officials) and school-based stakeholders (school proprietors, head teachers, teachers, local community leaders and parents). The analysis was guided by the research questions: How do stakeholders understand the implementation of PPP policy in USE delivery in Uganda? The literature on policy implementation emphasises the construct of understanding (Singh et al., 2014), without which the success of the policy is suppressed. This is consistent with Martinez et al. (2014)"s advocacy for sensitizing all stakeholders to the nature of PPPs, why they are needed and their possibilities and limitations in order to mitigate misunderstandings and inherent biases, which commonly cause information asymmetry and introduce barriers in their implementation. This confirms that stakeholders" understanding is an inevitable tool for successfully doing policy (Maguire et al., 2012).

Understanding of the origin of the policy
Most stakeholders perceived the origin of PPP policy in USE from the perspective of the government"s need to increase access to affordable USE, the problem of excess demand caused by the influx of learners into USE amidst insufficient capacity in public secondary schools, and budgetary constraints. These understandings are presented under the following emerging sub-themes.

Increasing access to affordable USE
Under this sub-theme, the findings indicate that most stakeholders expressed their understanding of the PPP policy in USE as having emerged from the government"s objective of increasing access to affordable USE for all children. In this regard, an inspector of schools viewed this policy as an alternative government strategy to increase access to EFA. She explained: … The major goal of this policy was really ensuring USE access for all; and since access for all could not be achieved in the government-aided schools, then the private schools were a second window through which increased access could be achieved. (Interviewee SH11) Likewise, one parents" representative from partnership school B perceived the policy to have arisen from the government"s need to achieve equitable access to USE. He stressed: About that policy, my understanding is that it was purposely meant to give equal opportunity to all the children, both the rich and the poor… So with the introduction of this policy…many of the children from poor families are also able to access secondary education… (Interviewee SH21) The above excerpts reveal that both government and school-based stakeholders share the understanding that the PPP policy originated from the government"s need to increase equitable access to affordable USE for all. This is consistent with the extant literature (LaRocque, 2009;Patrinos et al., 2009;Srivastava, 2010;Verger, 2011;MoES, 2013MoES, , 2014Mahmood, 2013;Mgaiwa and Poncian, 2016), which points out that ePPPs mainly arose from the need to expand access to quality public The preceding extracts indicate that most stakeholders perceived the introduction of PPP in USE as a policy response to the excess demand for USE by an influx of learners amidst the inadequacy and capacity limitations of government-aided secondary schools. This evidence is consistent with some scholarship (LaRocque, 2009;Mahmood, 2013), which affirms that PPPs in education resulted from excess demand for affordable public education services outstripping the capacity of public sector schools to satisfy it alone. This corroborates the central philosophy of PPPs that all organisations have strengths, but no organisation has all the strengths required to do everything alone (Reim, 2009;Rotter and Ozbek, 2010). Moreover, Patrinos et al. (2009) argue that, where the demand for education exceeds its supply owing to limited capacity and limited public funds, ePPPs become a suitable and cheaper policy intervention.

Budgetary constraints as the driver of PPP in USE delivery
Some stakeholders also perceived this policy as having resulted from the public budgetary constraints which the government experienced in USE delivery. One district administrator viewed this PPP policy as a strategic way of bridging the gaps caused by public resource constraints in education service delivery. The aforementioned stakeholders" perspectives seem to suggest that the PPP policy in USE arose from public resource constraints in USE delivery. This finding is consistent with that of Reeves (2013), who identifies fiscal resource pressures as the major trigger for the adoption of PPPs in the country"s public sector service delivery engagements. This concurs with Patrinos et al.
"s and the World Bank (2011)"s assertions that the scarcity of resources for public services delivery is the key justification for the adoption of PPPs in most developing countries.

Stakeholders' understanding of PPP policy guidelines
This sub-theme provides stakeholders" understandings of the implementation of PPP policy guidelines in USE (MoES, 2013). This study uncovered varied understandings of the policy guidelines among stakeholders and the existence of stakeholders who implemented a policy they did not clearly understand, which was mainly attributed to an imbalance in policy information access. For instance, when asked about her understanding of the policy guidelines, one teacher from partnership school C expressed lack of awareness of the guidelines for the policy she implemented. She attributed her lack of awareness to not being alerted to the existence of the guidelines. She remarked: Expressing a similar sentiment, one local community leader noted that "we would be able to know [understand] the policy if we were greatly involved in its implementation…but they didn"t inform us about it..." This confirms lack of a clear understanding of the policy guidelines by most school-based stakeholders, because they were neither informed and sensitised regarding the policy nor engaged in its implementation. This finding corroborates Higham and Yeoman (2009)"s affirmation that the degree of stakeholders" engagement in the partnership policy has a great influence on their understanding of it. On the contrary, most governmentbased stakeholders seemed to have a clear understanding of the PPP policy guidelines for USE delivery. In this regard, one local government official provided his insights into the policy process as follows: The above responses indicate that most government stakeholders exhibit clear and high levels of understanding the policy guidelines (as contained in MoES, 2013; PPP Policy Act for Republic of Uganda, 2015) compared to school-based stakeholders owing to their technical roles, work experience and great involvement in the entire policy implementation process. Yet most school-based stakeholders, who were neither informed nor sensitised regarding the policy guidelines, implemented a policy which they did not clearly understand. This finding seems to suggest that in certain contexts, some public policies may be implemented by stakeholders without having adequate policy information and a clear understanding of the policy guidelines. This finding partly bears out those of Martinez et al. (2014), who explored similar cases of stakeholders" misunderstandings and misperceptions of the PPPs in infrastructure in the USA. Such a context-specific anomaly therefore contradicts Maguire et al. (2012)"s assertion that the critical factor in policy implementation is sensemaking (understanding) by policy actors through effective communication. This revelation seems to suggest that variations in stakeholders" understanding of the PPP policy in USE were mainly informed and influenced by the differences in the contexts in which they implemented it (Blignaut, 2008;Palmer and Rangel, 2011).

Stakeholders' understanding of the PPP policy success in USE delivery
In this study, stakeholders perceived policy success based on the extent to which its objectives (MoES, 2013) of increasing access, quality and equity in USE delivery were achieved. Though most stakeholders shared the understanding that the PPP policy had partly succeeded in increasing access to USE through improved student enrolment, they disagreed and varied in their understanding of overall policy success with regard to issues of equity and quality in USE delivery. In this study, while most school-based stakeholders perceived the policy as unsuccessful (owing to the undermined equity and compromised quality of USE outcomes in most PPP schools), the majority of government-based stakeholders perceived it as successful because of increased student enrolment and access to USE. In the light of this, one government stakeholder said that "…the policy has been a success story because at least it has been able to bring more children on board to access secondary education and fill up gaps where government was not able to do…" Another government official noted that despite the existence of some gaps in the quality of USE, the policy was partly successful because of its unintended positive consequences for learners through increased access to USE. She elaborated:

…in terms of access, there is a big achievement. The number of children who have accessed secondary
Twinomuhwezi and Herman 139 education since the beginning of USE in the community has been a great success. …these children having the four years at school, even if they don't excel, but they definitely have grown physically and mentally, and they've been exposed more. So they are more civilised citizens than those who didn't have that chance. But in terms of the quality of results, we still have a big gap. (Interview SH11) In the same vein, another government stakeholder conceded that though access to USE improved, its quality had been compromised through the teaching and learning processes. He noted:

…while access has increased, quality of education has not really improved that much. … So it is an indication that quality is somehow compromised in terms of quality of teaching and learning…but we are saying that while access has been achieved largely, equity is not clearly achieved because quality is not across the board. Quality is an issue that we are grappling with as a Ministry… (Interviewee SH7)
In spite of the compromised quality of USE outcomes, most government-based stakeholders perceived the policy as successful owing to its positive impacts, including increased access to USE, reduced number of teenage pregnancies, a decline in dropout rates and absenteeism, attitude change, improved civilisation/exposure, skills development and acquisition, and the mental growth of learners. These findings are consistent with MoESTS (2014) and Gibson and Davies (2008)" argument that ePPPs not only impact positively on access to and quality of education but also on school enrollment, class attendance and attitude and behavioural change among students. Nevertheless, most school-based stakeholders and academics perceived the policy as unsuccessful owing to its failure to ensure equity and quality in USE delivery. In this regard, one school-based stakeholder pointed out that the policy was not successful as it did not meet its expectations. She emphatically said: No; and a strong NO! The policy has not created any great impact as we expected. We would expect a private partnership school to have improved in everything: its academic standards, its infrastructural development, and its retention of staff and even students but it is not the case. (Interviewee 16) In a similar vein, one academic perceived the policy as unsuccessful because both learner performance and the quality of USE or teaching received in PPP schools was lower than in non-PPP schools. He observed:

This policy promoted access which is a good thing but exacerbated inequity and quality among learners… when you compare what is happening in [partnership] private schools to other better USE schools, they are not getting the same quality education. So, PPP in USE is exacerbating the situation between the rich and poor, between the urban and rural… We know that equity has not been achieved and access is still limited, because most of the PPP schools are not in hard-to-reach areas… (Interviewee SH10)
The preceding insights indicate that most school-based stakeholders and academics judged the policy as unsuccessful mainly because of equity concerns and the low quality of USE delivered in most PPP schools. They perceived policy success based on the extent to which its goals of access, quality and equity were all achieved. Besides the compromised teaching-learning process, most stakeholders attributed the low quality of USE to inadequate capacity and resources as well as lack of regular school supervision of PPP schools. For instance, one stakeholder noted "…the quality of USE has been compromised by lack of constant and regular supervision mainly by government, yet every district has an inspector of schools". In this study, most non-state actors considered the success of the ePPP policy in connection with its USE quality and equity outcomes . Moreover, USAID (2008) points out that many countries have now shifted their focus away from the objective of access to that of quality in education. Likewise, Kaboru (2012:247) opines that "increased coverage [access] is highly desirable but insufficient unless it is accompanied by an improvement in the quality of public services". This understanding of policy success is consistent with Hodge and Greve (2011:11)"s assertion that the "PPP policy is deemed successful if its objectives are met and desired outputs are achieved". Based on their understandings of the policy realities, stakeholders suggested the following key success factors for its implementation.

Stakeholders' perceptions of the CSFs for PPP policy in USE
This section provides what stakeholders perceived as necessary factors for the success of PPP in USE delivery. It sought responses to the research question: What do stakeholders perceive as CSFs for the implementation of the PPP policy in USE in Uganda? These factors emerged from stakeholders" lived experiences, understanding and perceptions of this policy. The stakeholders" key insights into CSFs were categorised and discussed under the following emerging sub-themes.

Selecting suitable partners with adequate capacity
In this study, the findings reveal that most selected PPP schools experienced understaffing and resource inadequacies.
Most stakeholders consequently suggested the selection of suitable partnership private schools with adequate capacity in terms of quality facilities, quality teachers and monetary resources to provide quality USE outcomes and better school management. In view of this, one government stakeholder commented: The above response indicates that the government should select private schools with adequate quality staff, a well-facilitated school environment and good management in order to effectively deliver USE through the PPP policy. This finding is consistent with Patrinos et al. (2009)"s assertion that the capacity of a contracting agency/partner is paramount for the successful implementation of ePPPs. Moreover, Mgaiwa and Poncian (2016) argue that adequate quality of teaching staff is a key determinant of quality education and its outcomes through ePPPs. This finding on the selection of PPP partners with adequate capacity is consistent with those of most non-education PPP studies (Chan et al., 2010;Abeer et al., 2011;Cheung et al., 2012;Ismail, 2013; in which it is considered as a KSF for PPP projects. However, one stakeholder revealed that the selection of partnership schools was not fully based on the principle of adequate capacity owing to the political influence and the urgency with which the policy was incorporated into the USE agenda in Uganda. He elaborated: This finding supports Srivastava (2010)"s assertion that political influence is a factor in the choice of PPPs and partners despite their inability to provide quality services. This undue political influence in the selection of PPP partners can be regulated by establishing strong autonomous PPP units or structures.

Regular monitoring and supervision of PPP implementation
Having experienced lack of follow-up on partnership schools by the government as regulator of the PPP policy, most stakeholders suggested the need for regular monitoring and supervision of its implementation by enhancing the existing capacity and functional structures. In this regard, one government-based stakeholder suggested: There is need for strengthening the supervisory function, because the department responsible for that doesn't have the whole capacity to go to these schools regularly as possible… So, the supervisory function should be improved from the local community level structures. (MoES-Government official 1) Likewise, some school-based stakeholders suggested the establishment of a routine policy monitoring and evaluation mechanism to ensure stakeholders" adherence to its terms and conditions: … The government should provide a mechanism for monitoring and supervising the policy to ensure that the schools adhere to contents of the partnership agreement between them and the government…The other issue that government needs to ensure is routine monitoring of the programme in all schools wherever they are so that there is no diversion of resources... (School head teacher) The above findings, which suggest participatory regular monitoring and supervision of PPP implementation activities through local community structures, agree with the PPP scholarship (Latham, 2009;Hodge and Greve, 2011;Aslam et al., 2017). This acknowledges the need to strengthen the monitoring and regulatory capacity of government mechanisms while instituting indicators and quality assurance standards against which PPP partners and policy success are monitored and evaluated for compliance purposes. Similarly, the preceding findings and PPP literature (Jamali, 2004;Mahmood, 2013;Verger and Moschetti, 2016) posit that the government should set clear standards and build adequate capacity and mechanisms to ensure regular monitoring of ePPP policy to ensure its success.

Transparent regulatory and accountability enforcement mechanisms
This study established that the PPP funds were misappropriated and abused by some stakeholders owing to lack of transparent accountability structures, yet they continued receiving the funds. Based on such adverse lived experiences, most concerned stakeholders proposed that a transparent accountability framework be The above findings suggest that government should enhance the enforcement of accountability measures and internal controls in partnership schools to ensure ePPP policy success. These findings are consistent with the opinions of other PPP proponents (Forrer et al., 2010;Srivastava, 2010;Brinkerhoff and Brinkerhoff, 2011;Heald and Georgiou, 2011) on the need for transparent accountability mechanisms with clearly defined procedures against which PPP partners can be evaluated and held accountable. This proposal also corroborates the views expressed by UNICEF (2011) and UNESCO (2017) that if all stakeholders were held accountable for their roles and actions, ePPPs would improve the efficiency and quality of education service delivery. Besides, an NGO-based educator suggested the tagging of PPP funding to compliance and performance levels to enhance accountability by partnership schools in Uganda (Mahmood, 2013).

Stakeholders' commitment to partnership roles
The findings of this study reveal that unless stakeholders in the PPP policy in USE are committed to their partnership roles, its implementation will continue to be a challenge. In this regard, one concerned government stakeholder pointed out that this commitment is a critical factor for successful policy implementation. He stressed: We should get to know that this education policy is a collective responsibility… each and every stakeholder should do his or her roles. Government must increase funding, parents must provide scholastic materials for their children…and there should be timely release of funds, of course by the central government; and even the schools must account for the money on time and they must stop wasteful expenditures. ...each and every stakeholder should walk the talk, and do what is expected… and we try to behave as professionals … (Local Government official 6) Another school-based stakeholder emphasised that stakeholders" commitment and devotion to their roles are the most important success factor for this policy. He noted:

Commitment summarises everything else in partnerships.
If the government is committed to release money in time and if we who are on the lower level are also committed, we can all do what we are supposed to do as clarified. The issue is commitment, dedication and knowing that we are doing this as a service to the country not to ourselves or to anybody else… (PPP school head teacher) The two preceding extracts suggest that all stakeholders in this policy should be committed to their respective roles with professional integrity. This is consistent with Ismail (2013)"s work which perceives commitment and responsibility as fundamental principles in PPPs to which all stakeholders should adhere in the interests of all. This policy recommendation was also in line with Freeman (1984)"s stakeholder principles of corporate legitimacy and agency, which stress that each party must fulfil its roles to serve the interests of all stakeholders for the benefit of all. Thus, the success of ePPPs is sustainable where partners are committed to performing their respective roles.

Regular policy review
It also became evident in this study that the contexts in which stakeholders implemented the PPP policy in USE had evolved over time. Yet the terms and conditions, as stipulated in the PPP policy guidelines, had remained relatively fixed, restrictive and partly irrelevant since 2007. For instance, the policy terms forbade partnership schools from imposing any extra cost beyond the fixed PPP subsidy (equivalent to US$12 per child per term) that they received from the government, yet its value had drastically fallen. Such a mismatch caused most stakeholders to advocate regular ePPP policy review to make the policy flexible and compatible with the evolving contexts. In this regard, one school-based stakeholder suggested: We need to amend the policy to match it with the current trends. The policy was adopted in 2000s, but over that period of time things are changing. …For me the most critical aspect is that we need to amend the programme [policy] to incorporate more of the views of the stakeholders at a school level so that it becomes publicly acceptable by people who are implementing it… (School head teacher) This view corroborates one government official"s remarks that "…the policy has remained static and yet the circumstances have kind of evolved; so, the policy needs to be fluid and continually be revised, which has not happened." Another government official reiterated: We need to review the policy guidelines and make them much more binding and comprehensive by introducing aspects of quality. Because the guidelines are just broad, they are looking at mainly inputs but we are not tagging this partnership policy to the outcomes. So, the guidelines should have also come up with clear targets.
[If] there are no targets; how do you hold these schools accountable? (MoES official 4) The aforesaid policy narratives and recommendations conform to Forrer et al."s (2010) perception that regular policy review is crucial to successful policy implementation. This makes the policy and its guidelines clearer, simpler, flexible and more binding for compatibility within the evolving policy environment. This finding concurs with Mgaiwa and Poncian (2016)"s assertion that education policy review is a corrective measure for the successful implementation of education PPPs. A flexible ePPP policy creates a better fit between supply and demand for education services; otherwise, private actors would be pushed to illegitimately operate outside its set guidelines (Latham, 2009).

Timely provision of sufficient funding
Most stakeholders unanimously suggested that timely provision of sufficient funding to the PPP schools is a CSF for the effective implementation of the PPP policy in USE. In this regard, one school-based stakeholder proposed that the government should both increase the PPP subsidy and consider taking up the payment of teachers" salaries in partnership schools. He commented: I think government should increase funding to these schools. And then, apart from the funding they give per child, government should look at the tradeoff in terms of teachers' recruitment; they could probably not give the money directly to the school but they could be paying teachers in PPP schools…because these private schools, particularly in the villages, find it difficult to recruit science teachers … (PPP school teacher) One parent echoed: I think government should increase the funding… that money which schools receive is little…if the government could as a partnership take up some teachers' pay and be incorporated in the government salary scheme; I think that would help these PPP schools in delivery of quality USE. (Parents' representative) The preceding findings suggest the need for adequate funding to PPP schools and improved pay for teachers for quality USE service delivery. These findings support Brinkerhoff and Brinkerhoff (2011)"s argument that, to ensure quality public services through PPP, the governments of poor countries should retain the responsibility of financing the private partners. Education systems with high public funding levels have better performance outcomes (Verger, 2011).These viewpoints echo Srivastava (2011)"s argument that additional funding for ePPPs is necessary for improved quality of schooling, mainly in poorer communities with a large influx of learners. Srivastava"s argument supports one stakeholder"s view that "…you cannot have good quality of education without funding." This implies that quality of education is a direct function of funding. Thus, the quality of USE is a monetary issue, because the success of the PPP policy in its delivery will greatly depend on adequate funding to the better managed PPP schools.

Regular policy communication and sensitisation of stakeholders
Finally, this study reveals that while some stakeholders clearly understood the PPP policy in USE that they implemented, others did not. Some stakeholders attributed the variations in understanding the policy to lack of sensitisation and regular communication on policy, limited access to policy information and language barrier in policy communication. Therefore, most stakeholders suggested that stakeholders" awareness of the policy be enhanced through regular sensitisation and policy communication using appropriate channels and forums. In this regard, one stakeholder said: I think we need to put some extra effort in terms of sensitising the masses but specifically in making them understand their roles in the whole programme, because every stakeholder has to know his/her role…to make this programme successful. (School head teacher) Another stakeholder added: "But I think we need to have The above stakeholders" perceptions suggest the need to enhance stakeholders" awareness of the PPP policy through mass sensitisation, regular communication and dialogue. As OECD (2012) affirms, the stakeholders emphasize that the simplicity and clarity of policy communication packages are the best principle for PPP success. These findings agree with extant scholarship (Reim, 2009;Hodge and Greve, 2011;Reeves, 2013) on non-ePPPs, which emphasise regularity, consistency and clarity in communication through appropriate strategies and channels. Thus, to reduce policy information asymmetry, stakeholders should regularly communicate effectively to one another through appropriate policy packages and channels for PPP success. This is because sensitisation, communication and feedback ensure clarity regarding PPP policy, which enhances mutual trust and commitment among partners in its implementation.

Policy lessons and recommendations
The following key lessons learnt and recommendations are intended as practical future policy considerations for PPP stakeholders: (i) An effective policy communication structure should be established to promote stakeholders" awareness of the policy and reduce information asymmetry.
(ii) The government should legitimately select committed partnership schools with adequate capacity to deliver quality USE services.
(iii) Sufficient funding should be disbursed punctually to better managed partnership schools.
(iv) Transparent accountability mechanisms should be established against which PPP partners should be evaluated and held accountable.
(v) An enforcement and regulatory body (PPP unit) with adequate capacity and autonomy should be established to monitor and evaluate ePPPs regularly.
(vi) Performance standards and targets must be incorporated into the ePPP policy for quality assurance in USE delivery.
(vii) The PPP in USE should be regularly reviewed for compatibility and flexibility within the evolving policy environment.
(viii) The PPP policy in USE should not be seen as a "one-size-fits-all policy" owing to the different contexts in which partnership schools operate.

Conclusion
This study has contributed to understanding stakeholders" perceptions of the PPP realities and its CSFs in USE delivery in the Ugandan context. It has established that the PPP policy contexts which stakeholders experience seem to influence their perceptions and understandings of it. The study reveals that some stakeholders can implement the policy without clearly understanding its details owing to the contextspecific differences in which they implement it. Besides, this phenomenological interpretive oriented study has made a methodological contribution to narrative policy analysis research in the field of PPPs. It has revealed that the success of the ePPP policy depends, as its CSFs, on stakeholders" commitment to their roles, regular policy review, sufficient funding, selecting partners with adequate capacity, regular policy monitoring and supervision, transparent accountability mechanisms, regular policy communication and sensitisation of stakeholders. Besides, it has confirmed that certain CSFs for ePPPs and for non-education PPPs are similar possibly owing to some similarities in the nature/type of contracts, implementation experiences, challenges and contexts. These findings will inform appropriate policy reforms and best practices for the successful implementation of ePPPs in Uganda and other similar contexts worldwide. This study contributes enormously to understanding the country-and context-specific CSFs of ePPPs that may facilitate engagement in further policy debates on what PPPs may entail in other global contexts. Finally, further research should be conducted to explore ePPP opportunities for ensuring social distancing and resilience in equitable access to quality EFA in the Covid-19 era.