'Writing budgets for meetings and teas?': a multitheoretical analysis of intragovernmental coordination for multisectoral action for health in Uganda.

INTRODUCTION: Coordination across policy domains and among government agencies is considered critical for addressing complex challenges such as inequities, urbanisation and climate change. However, the factors influencing coordination among government entities in low-income and middle-income countries are not well known. Although theory building is well suited to explain complex social phenomena, theory-based health policy and systems studies are limited. This paper examined the factors influencing coordination among government entities at the central government level in Uganda.
METHODS: This theory-based case study used a qualitative approach. Primary data were collected through 26 national-level key informant interviews supplemented with a review of 6 national strategic and policy documents. Data were analysed abductively using a multitheoretical framework combining the transaction cost economics theory, principal-agent theory, resource dependence theory and political economy perspective.
RESULTS: Complex and dynamic interactions among different factors, both internal and external to the government, were found. Interdependencies, coordination costs, non-aligned interests, and institutional and ideational aspects were crucial factors. The power dynamics within the bureaucratic structures and the agency of the coordinated entities influence the effectiveness of coordination efforts. New public management principles promoted in the 1990s by donor institutional strengthening projects (characterised by agencification and setting up of independent agencies to circumvent ineffective big line ministries) created further fragmentation within the government. The donors and international agendas were occasionally supportive but sometimes counterintuitive to national coordination efforts.
CONCLUSION: The multitheoretical framework derives a deep analysis of the factors that influence organisational decision-making to coordinate with others or not. Achieving intragovernmental coordination requires more time and resources to guide the software aspects of institutional change-articulating a shared vision on coordination across government. Shaping incentives to align interests, managing coordination costs and navigating historical-institutional contexts are critical. Countervailing political actions and power dynamics should be judiciously navigated. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Despite the evidence that applying theory to research helps understand complex social phenomena better, theory-based studies on coordination of multisectoral action for health in low-income and middle-income countries (LMICs) are limited.

Empirical theory building stems from the high-income countries, but this may not be suited for LMICs.

This paper reports a theory-based case study on the factors that influence multisectoral coordination at the central government level in a low-income setting (Uganda).

Significant coordination problems such as duplicated mandates, incoherent policy actions and fragmented structures arise from interactions among different structural–institutional and operational factors, internal and external to the government apparatus.

Key influential factors included interdependencies, coordination costs, non-aligned interests, institutional and ideational aspects such as the logic of traditional bureaucracy specialisation enhanced by new public management principles and ethos.

Applying a multitheoretical framework derives a deep analysis of factors that influence organisational decision-making on whether and how to coordinate with others.

Better application of theory can enhance practice by providing lenses to anticipate coordination challenges and inform policy for multisectoral action for health.

Introduction

Achieving coordination of policy and practice across departments and programmes is one of the major problems facing governments in both low-income and middle-income countries (LMICs) and high-income countries (HICs).1 Coordination is a complex and multidimensional phenomenon with internal–external and horizontal–vertical dimensions.2 3 External coordination concerns the relationship between the government with non-state entities, while internal coordination pertains to intragovernmental aspects. This study focuses on intragovernmental coordination, defined as the process by which public organisations (ministries, departments and agencies (MDAs)) ensure ‘that their actions take into consideration the activities, resources and outcomes of other government organisations’1 (p9). Despite the focus on internal coordination, we acknowledge that external coordination bears on it in reality. Vertical coordination concerns different hierarchical levels, while horizontal coordination pertains to the same hierarchical levels. For the central government, internal coordination refers to the coordination among and across MDAs.

With a few exceptions (eg, Shankardass et al),4 the vast corpus of public health research focuses on external coordination between the public sector and non-public organisations such as academia, civil society organisations and donors.5–7 Collaboration between government and non-state actors (NSAs) is well captured in scholarship on governance and related concepts such as multisectoral governance, collaborative governance or network governance.8 9 Vertical (multilevel) coordination between the central and subnational levels is covered in the literature on decentralisation.10 11 However, there remains a lack of information on how internal coordination within the central governments, especially in LMICs, can be enhanced to advance health aspirations and social objectives.12 The central government level deserves studying as it is meant to perform core functions such as policy-making and supporting the subnational level governments.13 The functioning of the central government bears down on the overall performance of government.

In LMICs, inadequate coordination has been attributed to several factors such as weak institutions, limited resources, over-reliance on external actors and limited capacity.9 14 15 The central government represents a complex entity with multiple interorganisational relationships (IORs) within a dynamic and complex environment of interacting factors and actors. In government, various MDAs have primary mandates that shape interests and incentives for or against multisectoral action (MSA). In addition, there are often powerful agencies with constitutional powers to steer some government-wide processes such as budgeting, strategic planning and civil service management. The bureaucratic setup entails power dynamics as entities contest over mandates, ideas, scarce resources and diverse interests operating within multiple layered institutional settings.4 16–18 Uncertainty due to environmental dynamics and actors’ self-interest and opportunism (rent seeking) discourages actions across sectoral and organisational boundaries.15 Other issues of concern include transaction costs related to contracting and monitoring, alignment of political aims, development of shared vision, cross-organisational learning between implementing agencies, the control over resources and the management of interdependencies. All these factors cannot be understood by linear logic or a singular theory. They would benefit from relational thinking and analysis.16 19

Public sector reforms and coordination in Uganda

The Ugandan public sector underwent several reforms over the last three decades that have affected intragovernmental coordination. The late 1980s coincided with post-war recovery led by President Museveni and his National Resistance Movement government. This coincided with structural adjustment programmes and efforts toward leaner government inspired by the Bamako Initiative and driven by the Washington Consensus.20 Strong donor influence on national policy is well documented.21 Since the 1990s, rapid decentralisation has been marked by the creation of numerous local government units.22 In addition, several autonomous agencies have been created to perform specific functions, such as the Uganda Revenue Authority (for tax management) and Uganda National Road Authority (transport infrastructure development). However, the extensive decentralisation and intense agencification characterised by duplication mandates, inefficiencies and capacity gaps have been cited to significantly constrain intragovernmental coordination in the country.23

Several structures and systems have been established to perform coordination functions in government. According to the 1995 constitution, Uganda is a republic led by the president supported by an executive of ministers constituting the Cabinet. Several ministries such as the Office of the Prime Minister (OPM), Finance and Public Service are mandated to coordinate cross-cutting government functions. The parliament was introduced to make laws and hold the executive accountable. The judiciary was created as the third pillar of government.24 These government systems are expected to work in a complementary and synergistic manner, but this is not the case in reality.23 25

Coordination has also been pursued around national development aspirations espoused in national strategic plans and documents such as the poverty reduction strategy papers, Plan for Modernisation of Agriculture and poverty eradication action plan. National development efforts, from 2010, converged around the comprehensive national development planning framework comprising the national Vision 2040 and a series of 5-year national and sectoral development plans.26 Other government-wide efforts to enhance coordination have included recentralisation, adopting sector-wide approaches, creating sector working groups (SWGs) and functionalising these coordination structures and plans.23

Theory building

The study was inspired by the fact that although theory building is well suited to explain complex social phenomena,27 28 the application of theory in health policy and systems research (HPSR) studies is limited.29 30 Reeves et al31 affirm that ‘theories give researchers different ‘lenses’ through which to look at complicated problems and social issues, focusing their attention on different aspects of the data and providing a framework within which to conduct their analysis’ (p1). Some recent HPSR studies have applied theories to varying public health problems and in multiple ways. For example, theory has been used to derive and describe research problems.32 In other instances, theory-guided data analysis has been used to interprete and explain data.33 34 There is also an interest in the number of theories used in a given research enterprise.35 There are instances when one theory has been used to interprete empirical data in LMICs-based case studies.33 36 Conversely, some HPSR scholars have used more than one theory in their inquiry.37 The multitheoretical approach derives richer and deeper analyses by drawing on the different theoretical world views and propositions, thereby overcoming the potential slanted analysis associated with no or a single theoretical perspective.35 38–40 Indeed, combining theories has been a hallmark of (critical) realist inquiry applied in recent public health research.4 41–44

On the contrary, empirical theory-based studies on (intragovernmental) coordination of MSA for health in LMICs are limited. Most theory building is still mainly HIC based, obfuscating certain aspects of public administration common to both HICs and LMICs but may be more apparent in LMICs. Nevertheless, theoretical application has been considered in some recent review articles. For example, Bennett et al14 applied the political economy (PE) perspective to examine factors influencing multisectoral collaboration in LMICs. Ssennyonjo et al12 explored the potential of four social science theories to inform intragovernmental coordination efforts in LMICs. Shankardass et al4 provided a theory-based framework for examining coordinated government action based on systems’ theory. Generally, the potential of theories in informing practice is yet to be optimised.30

Using the case of the coordination at the central government in Uganda, this study explored and analysed enablers and constraints to achieving intragovernmental coordination using a multitheoretical approach. The research question guiding this research was: What factors influence coordination among MDAs at the central government level to advance multisectoral action for health in Uganda, how and why? The rest of the paper is organised as follows. First, the Methods section highlights data collection procedures and the application of a multitheoretical framework to data analysis. Then, the Results and Discussion sections follow. Finally, the Conclusions section is provided.

Methods

Study design

The study adopted a theory-driven case study design to analyse the factors that shape intragovernmental coordination in Uganda.45 46 Uganda was a viable study setting because it bears similarities with other LMICs regarding the institutional set-up in government. An executive branch of government is divided into specialised units, that is, MDAs.25 The ministries have ministers as their political heads and are supported by a bureaucracy of technical officials. Being a low-income country, resources available to the government are limited, and the country has a heavy donor reliance.47 Accordingly, coordination issues and priorities are shaped heavily by global and regional agendas and actors.48 The study mainly focused on the 2015–2020 period, which coincided with the National Development Plan (NDP) II, the Health Sector Development Plan and at the international level, the sustainable development goals (SDGs).

The study used document review and key informants’ interviews with governmental officials and NSAs within and outside the health sector. This paper draws from the first author’s PhD study looking at different aspects of multisectoral coordination for health in Uganda such as coordination instruments and the relationship between the Ministry of Health (MOH) and other MDAs.

Multitheoretical framework

The political scientist, Paul Cairney,35 points out three approaches to applying multiple theories: (1) identifying the most appropriate theory, (2) using the theories in a complementary manner and (3) synthesising several theories into a new one. According to Cairney,35 the choice of the multitheoretical approach depends on the envisaged benefits to the study. This study thus adopted the second form of multitheoretical approaches to derive complementary and, at times, competing explanations from the selected theories on why government organisations coordinate or not. The authors proposed a theoretical framework grounded in four theories that consider IORs as means to coordinate.12 49 The identified theories were transaction cost economics (TCE), principal–agent theory (PAT) or agency theory, resource dependence theory (RDT) and political economy (PE). They provide underlying complementary and, at times, competing explanations on why organisations coordinate or not and how they choose to do so. These theories all start from a focus on relationships between organisations working in a system (which can be a public sector, a government administration or government as a whole). The insights from the theoretical lenses were derived from the literature on IORs.49–51 Table 1 provides an overview of these theoretical insights.

Table 1

Multitheoretical framework for analysing drivers for and obstacles to IORs in government

Theories	Propositions	Main motivations for coordination	Coordination obstacles
TCE	The costs of the exchange relationship and the need to reduce these costs are the main drivers of coordination mechanisms	Efficiency	High transaction costs are associated with negotiating, monitoring and enforcing a contract Opportunism of actors is reinforced by uncertainty and environmental complexity because of the bounded rationality of the options and the ‘small number’ of problem/uncontestable mandates of actors
PAT	Coordination of multisectoral actions requires incentives/sanctions to encourage organisational motivation. Principals apply external controls and contracts to create incentives for agents to coordinate	Power of incentives	Bounded rationality and opportunism of actors and complexity in the performance environment Uncertainty over outcomes Conflicting goals and interests between the principal and the agent Different attitudes toward risk between the principal and the agent
RDT	Organisations seek to find mechanisms to ensure the smooth and predictable flow of external resources from those who control them. The coordination is driven by acknowledgement of interdependence that arise because each actor possesses resources needed by another	Stability of resource flow Interdependence/reciprocity for mutual benefits	Limited acknowledgement of interdependencies Asymmetrical control logics promote competition instead of cooperation
PE perspectives	Coordination dynamics and mechanisms evolve because of a political (contested and negotiated) process Internal politics of coordination unfolds within the dynamic interactions between actors and context Actors are enmeshed within broader structures and complex power, and institutional and ideational contexts that condition their actions. Over time and when conditions are amenable to change, actors’ agency can influence structures and create new structures	Necessity in response to regulatory instruments Legitimacy in response to normative elements Path dependence and institutional structures Power structures and motivations of those that hold vital power for needed actions (resources, positional, knowledge reverence, etc)	Conflicting institutional logics and interests promote organisational silos Multiplicity of power centres arising from institutions, ideas and agents History and normative values for action or inaction

Source: Ssennyonjo et al.

IORs, interorganisational relationships; PAT, principal–agent theory; PE, political economy; RDT, resource dependency theory; TCE, transaction cost economics.

The theories propose different views on context conditions and social mechanisms for initiating and implementing coordinative relationships by (public sector) organisations. For instance, the TCE considers the transaction cost related to exchange as the main driver for choosing (how) to coordinate. From the other theories, understanding the nature of interdependencies and distribution of resources (RDT), the existing principal–agent relationships (agency theory) and the power dynamics and control over resources in a given historical political context (PE) is critical for a multisectoral action.

Data collection methods

Interviews and respondent selection

A total of 26 key informants categorised as seven MOH officials (1–7), 14 non-health MDA government officials (1–14) and 5 NSA (1–5) were interviewed. The respondents comprised purposively selected national-level government officials and NSA from the health and non-health sectors as per table 2.

Table 2

Respondents per category

Respondent category	Organisational affiliation
Government	MOH	7
	Non-health MDAs	14
NSAs	Donors	2
	Academia	1
	Civil society organisation	1
	Non-government organisation	1
Total		26

MDAs, ministries, departments and agencies; MOH, Ministry of Health; NSAs, non-state actors.

The government officials were chosen based on their government positions and ranged from the senior officers to ministerial levels. The inclusion criteria were (a) being at the minimum a senior officer in the government system and (b) working in a department or agency with coordination functions for either the whole government or ministry or specific cross-cutting health policy issues such as maternal child health, nutrition, epidemics and non-communicable diseases. The NSA were included to gather their ‘outside’ perceptions of the workings of government. For non-state respondents, inclusion criteria were (a) being a national level actor and (b) having experiences of working directly with government as a member of technical working groups or sector governance structures such as the Health Policy Advisory Committee.

The extensive work experience ensured that the respondents could comment on institutional trajectories over a more extended period from different positions they had taken up (presumably) over the course of their careers. A list of key MDAs and potential respondents with their contacts was generated in consultation with the study team and key stakeholders in the MOH, OPM and the National Planning Authority (NPA).

Respondents were initially contacted through telephone calls and email, briefed about the study objectives and invited to participate. Informed consent was provided by the respondents. An interview appointment was booked based on the interviewee convenience.

Interview procedures

The key informant interviews were conducted in English by the first author (AS), assisted by two graduate-level research assistants. These interviews were conducted face to face between December 2019 and March 2020 at the respondent’s workplace. Most interviews (25/26) were audio-recorded and supplemented by field notes. Extensive notes were taken for the non-taped interview. Interview duration ranged from 20 min to 1 hour and 30 min, mainly dictated by the respondent’s time constraints. Priority questions were asked for very busy high-level officials. A semi-structured interview guide was developed for this study by the first author and reviewed by the other authors. The tools were adapted to the respondent category. They contained open-ended questions and probes based on theoretical perspectives and public administration literature on intragovernmental coordination. The tools focused on several questions, but this paper reports on the analysis of the questions related to motivations and barriers for multisectoral coordination.

Document review

The documents for review (ie, National Coordination Policy,52 sector development planning guidelines,53 sector development planning regulations 2018,54 the NDP II (2015–2020)55 and Vision 2040)26 were purposively selected as they represent the main strategic policy and planning documents for coordinating national development in Uganda over the 2015–2020 period. The National Coordination Policy was developed by the OPM. In contrast, the NPA developed the sector planning guidelines, NDP and Vision 2040. These national documents were supplemented by the evaluation report on the coordination function at OPM and SWGs.23 All the documents (except two obtained from interviewees) were readily available in public domains.

Data management, analysis and interpretation

The audio recordings were transcribed verbatim by two research assistants who participated in the data collection. The first author, assisted by two experienced postgraduate level researchers, read through the transcripts to ensure they captured all the information. Atlas Ti V.9, a qualitative data management software package, was used to code and analyse the data from the interviews, extensive notes for the non-recorded interview and documents.56

Data analysis was informed by the critical realist paradigm, which explores mechanisms that underlie empirical observations.19 57 We generally followed an abductive process as applied by Harris et al,43 where data and theory were used iteratively to interprete and deepen meanings of empirical data. The constructs from the multitheoretical framework informed the analytical themes. For the document review, full-text reading was done to orient the reviewers on the contents of the documents. This was followed by searching for text corresponding to the following keywords: ‘coordination, health, Ministry of health, institutional framework and structures’. Finally, the derived text excerpts were exported to Microsoft Excel for further analysis. For data from the interviews and extensive notes, pieces of text related to specific codes were highlighted and the code attached. Initial inductive open coding was followed by line-by-line coding of the data. After that, query reports were generated for the codes, followed by sorting and grouping texts with similar meanings into subthemes. Finally, the related subthemes were grouped into themes. We used the Consolidated Criteria for Reporting Qualitative Results58 to guide the study and report the findings. Illustrative quotes are used to exemplify and substantiate the subthemes. Analysing the data from both data collection methods was concurrent and iterative to boost the integration and validity of the findings.

Results

The section presents the factors that influence coordination among government entities at the central government level in Uganda over the period 2015–2020. The findings are organised according to the concepts from the theoretical frameworks guiding this study.

Coordination-related transaction costs

The TCE and agency theories predict that costs associated with coordination, including search costs, designing and enforcing contracts and performance monitoring, influence coordination decisions.49 From the interview narratives, we deduced two categories of coordination costs: (a) costs borne by the central agencies with coordination mandates and (b) those faced by the coordinating entities as summarised in box 1.

Shortages in financial and technical resources were framed as organisational capacity gaps.

Competing priorities of the coordinating agencies exacerbated the resource gaps.

Legal-institutional frameworks undermine efforts to respond to the gaps above.

Accountability procedures in government constrained efforts to negotiate collaborative efforts or mobilise resources to service the mandates.

Coordination was perceived to be costly and a risky endeavour associated with actual loss of benefits (such as resources and control) in an environment of competing public sector organisations.

Competition over priorities led to competition over resources during annual budgeting processes.

Tensions exist between demanding designated coordination budgets or making internal budget adjustments at MDAs.

Opportunity costs of coordinating with others

MDAs, ministries, departments and agencies.

Costs borne by the central agencies with mandates to coordinate others

In the Ugandan Central Government, some agencies are mandated to coordinate cross-cutting functions such as planning, budgeting, implementation monitoring and statistics (Ssennyonjo et al, under review). The costs borne by such agencies (eg, the Prime Minister’s Office) were considered a predominant constraint to coordination. These costs were often presented as financial and technical resources whose shortages were framed as organisational capacity gaps. The gaps pertained to the capabilities to map actors, specify roles, develop explicit contracts and monitor other entities (as agents).26 52 59 As a result, coordination duties reportedly outmatched the existing human resources pertaining to time and technical competencies.

Regarding OPM’s mandate to coordinate government business, one MDA official emphasised

‘The (OPM coordination) department is very understaffed. We are talking about eighteen sectors (and) about six people. So, you find one person is doing a lot and nothing’ (MDA-8).

The capacity gaps rendered these entities weak as principals mandated to coordinate strong agents and unable to enforce coordination requirements.

The coordinating agencies’ broad mandates and often competing priorities exacerbated the resource gaps. For example, the OPM as coordinator of government business was said to be distracted in non-coordination functions. Relatedly, the financial resources dedicated to coordination were considered not commensurate with the demands of coordinator roles and stakeholder expectations., This was partly attributed to the underappreciation of coordination costs. One NSA observed:

The (OPM) coordination department spends less than three billion annually to coordinate the whole government …. It is less than three billion to coordinate twenty-seven trillion. (NSA-1)

A non-MOH government official further emphasised:

I have to mention the issue of resources. People do not believe in coordination. (They say) ‘How can you keep writing budgets for meetings and teas?’. (Yet) you cannot call people (or a meeting) and not give them anything. Besides teas, real coordination needs money. (MDA-8)

Attempts to overcome resource gaps were reportedly further undermined by gaps in the legal-institutional framework, that is, inadequate description of roles, responsibilities and accountability relationships. For example, bureaucratic controls surrounding procurement and accountability procedures constrained operational level actors to effectively negotiate collaborative efforts or mobilise resources to service their mandates.

Opportunity costs of coordinating with others

Consistent with the conventional conceptualisation of transaction and agency costs related to the development and monitoring of contracts, coordination was perceived to be costly and a risky endeavour associated with actual loss of benefits (such as resources) in an environment of competing public sector organisations. In addition, risk-averse organisations reportedly harboured concerns about losing control, making such actors less enthusiastic about engaging in coordinated initiatives. According to a civil servant of the MOH:

Coordination is expensive not only in terms of direct costs to optimally coordinate people and institutions with different mandates but also in terms of the opportunity costs. People tend to look at coordination activities as taking them away from their core mandates and spending time at the interface between one sector and another. (MOH-1)

This dynamic was said to lead to fights over priorities during the annual government budgeting processes. Acts of self-preservation made MDAs introduce many new activities during the budgeting processes. One non-MOH official said:

It is all about protecting the budget, protecting your budget or the money that comes to your ministry. People think that when you accept to work together, you are ceding activities. When you cede activities to another agency, it means you are giving up money. (MDA-1)

Tensions between demanding designated coordination budgets from the Finance Ministry and making internal budget adjustments were reported. Reorientating organisational activities and budgets were deemed risky due to the foregone benefits accrued from attaining the core mandate’s objectives. The MDAs were said to often renegade on coordination commitments.

The resources are limited. [under such circumstances], people would concentrate on their core mandates, and you (calling meetings) will not be given a priority. (MOH-2)

Principal–agent linkages: (inter)dependences, interests and agent behaviour

Several principal–agent relationships were highlighted as influential to intragovernmental coordination. These included politicians–technocrats, donors-government and horizontal interministerial relationships. There was consensus that agency problems such as varied interests, hidden actions (moral hazard) and hidden information (adverse selection) characterised these relationships and undermined coordination. Box 2 summarises findings from this theme.

Agency problems such as varied interests, hidden actions (moral hazard) and hidden information (adverse selection) characterised the principal–agent relationships in government and undermined coordination.

Interdependence was considered to lead to either cooperative or competitive behaviours depending on whether the interests were perceived as mutual among the (would be) coordinating parties or not.

Collaborative efforts often arose from opportunistic motives to meet organisational interests such as fulfilling mandates and accessing resources owned by other agencies or ‘from the collective pool’.

Coordination opportunistically pursued to shift risks of potential policy failure to others.

Actors hide information and their true interests during planning until later during the implementation of interventions.

Self-interest is exacerbated by the institutional setting, where MDAs report their organisational performance only and not collective performance.

MDAs, ministries, departments and agencies.

Consistent with theoretical insights from TCE and agency theory, respondents attributed these problems to opportunistic behaviours, the difficulties in monitoring and managing these behaviours (eg, bounded rationality) and uncertainty.50 In line with RDT thinking, interdependence was considered a key driver for coordination. It was said to lead to either cooperative or competitive behaviours depending on whether the interests were perceived as mutual among the (would be) coordinating parties or not.

Our findings revealed that collaborative efforts often arose from opportunistic motives to meet organisational interests such as fulfilling mandates and accessing resources owned by other agencies or ‘from the collective pool’ An MOH official observed that:

Different sectors do observe that there is a comparative advantage they can exploit by working with other sectors. So, they do it (coordination) as a method of fulfilling their own objectives, and so that motivates them to collaborate and coordinate. (MOH-1)

A non-health government official agreed

People look at a multisectoral approach as an opportunity of synergies given the limited funding and being able to deliver your mandate or objectives with limited resources and using other people’s expertise. (MDA-5)

Risk aversion and blame games were also reported to characterise the pursuit of coordination. Interviews revealed some attempts to shift risks of potential policy failure to others. For example, the OPM was said to have asked the MOH to plan for refugee health after their own resources had been depleted.

The varied interests were said to lead to, at times, competition instead of cooperative behaviours. Several organisations were said to prefer ‘working in cocoons’ and not share resources with others to ‘maintain their flags’ and enhance their own reputation.

Consistent with PAT, some MDAs reportedly hide information and their true interests (exhibited adverse selection) during planning until later during the implementation of interventions. Several respondents remarked that such behaviours led to these instances of ‘holding up’ coordination (moral hazard) later during implementation. Two non-MOH officials argued that:

You have interests (but) you hide them. A policy comes (in place), and those interests continue playing out. They will completely just frustrate it. (MDA-1)

One of the main reasons [for limited coordination] is that people want to work in silos. Even if they bring them together, they can accept verbally. (MDA-2)

In addition, collaborative endeavours were said to disincentivise coordination by blurring organisational boundaries and creating accountability challenges. ‘We have seen some sectors attribute some outcomes to interventions that have not been primarily done by them’, one MOH official reported. These disincentives were perceived to be enhanced by an institutional setting, wherein MDAs report individual organisational performance only. On the other hand, reporting on success under a collective banner encouraged opportunistic MDAs to coordinate while discouraging the aggrieved ones.

The historical-institutional context and path dependency

In line with PE theory,60 61 path dependency was considered a vital attribute of the Ugandan government systems through which history shaped the prevailing and future actions and systems. Consistent with theory, interviews and reviewed policy documents23 53 55 revealed that institutions have both facilitative and constraining influences on intragovernmental coordination. Similarly, they are actively shaped and evolve as actors exercise their agency.62–64

Critical aspects of the institutional context in Uganda summarised in box 3 were (a) donor dependence, (b) agencification with organisational specialisation, (c) legal-institutional frameworks and structural power and (d) norms on multisectoral engagements. Informal institutions related to unsanctioned behaviours such as rent seeking and corruption were reportedly constraining forces.

Donors were reported as critical principals to the successive Ugandan governments that drove international norms and agendas for coordination over several decades.

Donor efforts embody narratives and framings of problems and solutions around which coordinated national-level efforts were rallied.

Donor initiatives carry norms and ideas of what matters or what is to be funded and the resources that influence how countries respond.

New Public Management reforms that resulted in vertical and horizontal specialisations led to multiple departments and agencies with different and, at a time, overlapping and conflicting mandates, systems, activities and funders.

Longstanding bureaucratic structures and the culture of silos in government were significant constraints.

Legal and policy frameworks were generally supportive of intragovernmental coordination.

Legal frameworks were inconsistent and maligned incentives.

Coordination structures were rife with power dynamics. For example, within Cabinet, some ministers are more powerful than others.

A culture of ‘untouchables’ in government frustrated efforts to work together.

Informal institutions related to unsanctioned behaviours such as rent seeking and corruption had constraining effects.

Rules and regulations were said to be only ‘on paper’ with limited enforcement.

Inadequate stakeholder mapping and engagement were significant features of interorganisational culture and barriers to developing a shared understanding within government.

Delegating junior staff instead of legitimate officials to intersectoral fora was a countervailing practice.

Different MDAs were had dominant professional groups with varying training, ‘languages’ and approaches to operations.

Norms and practices related to multisectoral engagements

MDAs, ministries, departments and agencies.

Donor dependence and international agenda influence

Donors were reported as critical principals to the successive Ugandan governments that drove international norms and agendas for coordination over several decades. The respondents cited several examples of coordination efforts being part of larger global development endeavours in the (health-related) areas of nutrition, universal health coverage and reproductive (maternal) and child health. There was consensus that these donor efforts embody narratives and framings of problems and solutions around which coordinated national-level efforts were being rallied. These initiatives were also noted to carry norms and ideas of what matters or what is to be funded as well as the resources that influence how countries respond. For instance, the SDGs were perceived as aligned to national coordination efforts in Uganda. A non-MOH official acknowledged that

Another good thing for coordination is that of the sustainable development goals because they entirely fit into our mandates. We have a call for all sectors to be contributing to them. So, when you call on them (MDAs) concerning SDGs, they always come and provide guidance. (MDA-8)

Donor funding was said to often catalyse MSA in the recipient countries by synergising effects of the government efforts.59 However, there were reports of antagonistic effects based on donors’ interests and the incentives these induce. One non-MOH official observed

But you know that (donor funding) is not sustainable. It will give you more mileage if you yourselves, your systems are strong. When they find you more organised, they give you more results in a short time. When you are more disorganised, then they disorganise you further with their own independent desires or directions. (MDA-8)

Agencification and culture of silos

The findings indicate that donor influence on intragovernmental coordination mainly manifested as the institutional architecture based on new public management principles promoted by Uganda’s development partners in the late 1990s. The resulting vertical and horizontal specialisations led to multiple departments and agencies with different and, at a time, overlapping and conflicting mandates, systems, activities and funders.

In addition, the longstanding bureaucratic structure and culture of silos in government were significant constraints to attaining multisectoral coordination. According to a non-MOH official:

It (coordination) is not an easy concept in our systems because of the way our sectors are structured. Our government delivery system has been structured that sectors deliver as single agencies. They deliver in a vertical manner, not both vertically and horizontally. If it is health (sector), they look at health alone. If it is water (sector), they look at water alone and do not consider how water is linked to health. (MDA-3)

The government’s financing and budgeting traditions reportedly reinforced incentives for perpetuating siloed government action. One non-MOH government official argued that:

It is not easy to break it (siloed approach) because this is how things have been for quite a long time. The way we have structured our budget promotes the silo approach because you have the ministry with its votes, and money goes to those votes. (MDA-1)

Formal versus informal institutional arrangements

Many respondents and documents23 26 52 lauded the existence of legal and policy frameworks supportive of intragovernmental coordination. Coordination was said to emanate from the necessity to comply with such legal imperatives. Central level agencies with coordination functions often used their structural power to bring together government agencies to coordinate. One non-MOH official provided an illustrative example:

Nutrition is a crosscutting issue. It is not only a health issue nor an agricultural issue, trade or an education issue … So, we took it up as National Planning Authority because it is a key development issue for the country. We coordinated planning to come up with a (multisectoral) (nutrition) action plan. (MDA-3)

Several limitations in the institutional context were identified. First, the legal frameworks were said to provide inconsistent and maligned incentives.23 For instance, accountability requirements of the Public Finance Management Act reportedly generated disincentives for joint action. Pooling and sharing resources were conceived to misconstrue collaborative efforts as misappropriation of resources while encouraging actual misaccounting of resources and results in other instances. An MOH official remarked that:

… unfortunately, that is the way the investigative arms of government look at it (cross-sectoral investments) … They tend to say you have misallocated resources. The other area is miss-accounting for interventions because some sectors may actually account for interventions that might have not primarily been done by them. (MOH-1)

Second, coordination was presented as a political process rife with power imbalances and infighting within government. Respondents noted the power dynamics within Cabinet, where some ministers were more powerful than others. The interviewees also reported a culture of ‘untouchables’ in government frustrating efforts to work together. Related to this were concerns about bureaucratic principles that designate the permanent secretary as the main spokesperson for the ministry, undermining the confidence with which government officials engage with other sectors.

Third, segments of respondents reported ongoing struggles to balance the power and incentives within formal and informal relationships and institutions. Many rules and regulations were said to be only ‘on paper’ with limited enforcement. The inconsistency between the ideal and practice was attributed partly to the actors exercising agency to actively undermine them and attain their respective interests, as confirmed by the non-MOH official:

Systems are in place, but you know, human beings beat the systems. The systems like the accountability mechanisms, codes of conduct are there. Everything is in place but just the people. In fact, like I always say, there is a lot of indiscipline and impunity across the whole field because you find people know the right thing but decide to do otherwise. (MDA-2)

Fourth, the interviews revealed that corruption, enhanced by weak enforcement of laws and a culture of rent seeking and poor accountability, led to interorganisational distrust. Respondents noted that coordination implies holding each party accountable. Yet, this need for ‘too much’ transparency was construed to have high transaction costs and risky in an institutional setting with counterproductive actions by some actors and instances of corruption. The hesitancy to collaborate was thus explained by the need for some actors to protect their corrupt cliques. One non-MOH official argued that:

Then the other thing is there is a lot of distrust, and this is actually motivated by corruption. You know there is a lot of corruption, and it is a very big problem. People feel that once they share information, (or) work together … some people might discover their loopholes of corruption. (MDA-2)

Norms and practices related to multisectoral engagements

The analysis revealed several features associated with multisectoral engagements. First, inadequate stakeholder mapping and engagement was exposed as a significant feature of interorganisational culture and a barrier to developing a shared understanding within government. This led to duplication of actions, for example, the Ministry of Water and Environment and the MOH were said to have parallel efforts to expand latrines and sanitation coverage. A few respondents noted limited acknowledgement of interdependencies. The interlinkages among MDAs were reportedly not evident nor well communicated. Multisectoral efforts were perceived as attempts of patronage and external control. An MOH official asserted that

These ministries ask, ‘why should the ministry of health come and guide us’. There is a communication problem … and I think if solved, working with other sectors will be much easier. (MOH-1)

Second, some respondents noted a countervailing practice of delegating junior staff instead of legitimate officials to intersectoral fora. This practice was attributed to the superiors (as principals) wanting to maintain control over these processes and not trusting their (competent) subordinates. The above practice was often interpreted by collaborating parties as a sign of disrespect and a gesture of disinterest in collaboration.

Third, different MDAs were reported to have dominant professional groups with varying training, ‘languages’ and approaches to operations. For instance, the MOH was reportedly predominantly occupied by clinical professional groups having a strong professional identity (grounded in a specific socialisation process during training). The lack of a common language frame and working ethos among civil servants at the central level was perceived to constrain interprofessional collaboration and working across sectors. According to one MOH official:

The other internal issue I can think of is the ministry of health is filled with doctors, and with the nature of their training, they tend to look inward. Sometimes, we find it very difficult to interface with the other sectors. (MOH-1)

Ideational factors

The interviews and policy documents23 26 52 55 revealed that framing problems and solutions did, and at times did not, support multisectoral collaboration. Considering development problems and solutions as crosscutting and requiring MSA was said to underlie many government and donor funding decisions. The example of nutrition was presented to demonstrate how ideas of interdependence successfully shaped multisectoral efforts in Uganda, leading to the adoption of a multisectoral Uganda Nutrition Action Plan.23 55 59 The ideational factors are summarised in box 4 and illuminated subsequently.

Disparities in understanding objectives being pursued.

Coordination considered a rolemandate of specific agencies and not cutting across levels and organisational boundaries.

Understanding that coordination is not a panacea and, at times, being counterintuitive.

Limited interest in information sharing due to political and technical reasons.

Political reasons were (a) protecting information on the organisation’s strategic interests and (b) power dynamics between political leadership and technical officers at MDAs.

Technical reasons included (a) inconsistences in organisational representatives to collaborative efforts, (b) limited opportunities to feedback to sectoral players, (c) fragmented information systems and (d) staff transfers in public service.

Limited evidence generation and use in government partly due to suboptimal efforts to evaluate (multisectoral) interventions.

Government as a learning organisation

MDAs, ministries, departments and agencies.

Varied understanding of coordination and its objectives

The findings revealed that the understanding of the objectives pursued under coordination varied. Coordination was mainly linked to the pursuit of efficiency and effectiveness toward common goals. An MOH official asserted that:

The market (government entities) poorly understands why coordination is important. Because of that, there is a very big problem that we have to surmount first. We have to make the market aware that coordinating actions reduces resource use and, in so doing it improves efficiency. (MOH-1)

There were also sentiments that coordination is not a panacea and could undermine other goals such as innovation and accountability. One NSA argued that:

Coordination or multisectoral work is not the answer to everything. Actually, the more you add, the more you will lose quality of detail …. So, you shouldn’t be obsessed so much with multisectoralism. You should also question when it is relevant. (NSA-4)

Our analysis revealed that the varied understanding of coordination had implications on MDAs’ perceptions of what was to be achieved by coordination, how and their respective contributions. Coordination was not considered to occur within and across all government levels, despite emphasis by some respondents that it should be ‘everybody’s responsibility’. According to one non-MOH official

The other issue is people don’t understand coordination. I have engaged several stakeholders, and there is a misconception that coordination is OPM. When you tell people ‘why don’t we do better coordination’, they will tell you ‘why do you want to do OPM mandate?’. I tell them that’s not true. Coordination is done at different levels. Even at my department, I coordinate … so, it’s not only OPM that brings people together to work together, no? So the problem is actually a very low understanding of coordination. (MDA-2)

(Dis)incentives for information exchange

Reinforcing the lack of shared understanding of coordination were inadequacies in information exchange reportedly arising from both political and technical factors. Politically, two factors were elicited. First, concealment of information was apparently inspired by strategic concerns that disclosing such information could compromise future organisational interests. This uncertainty and notions of distrust formed strong incentives for stringently guarding information. One non-MOH official argued that

Information sharing is done, and that’s a good thing. Though the information shared is selective. They (MDAs) don’t share it all. Some crucial information they hold to themselves so that they are the monopoly of the information. (MDA-2)

Second, the interviews indicated the differences among the interests and views of the political heads, on the one hand, and technical officers, on the other. Whereas the technical officers could agree to the multisectoral intervention, their political leaders would not necessarily do so. The malalignment of interests and perspectives personified power dynamics within MDAs and was said to undermine organisational commitment to collaboration. One non-MOH official narrated:

I have really seen this so much where the technical leadership agrees on something. The political leadership is saying, ‘whom did you consult?’. And I have learnt so much that the political economy plays a very big role in the multi-sectorial approach because technical persons sometimes do their work. Still, when it comes to binding the top leadership of these agencies, divergent ideas start coming in, and that will make everything collapse. (MDA-1)

The respondents linked inadequate information sharing to four technical reasons. First, inconsistencies in who participates in the multisectoral fora and variations in their respective competencies. Actors who participate in policy-making differ from those tasked with implementation because bureaucratic systems dictate that those who participate in the different policy development phases are different. Second, opportunities to feedback and update the sectoral players on the outcomes from multisectoral deliberations were limited. Failure to brief the sector players and broader governance structures within an MDA reportedly undermined the ability to follow through with multisectoral commitments/action points.

Third, fragmented information systems across and within sectors undermined information exchange. For example, one non-MOH official noted that ‘some projections on key indicators used by MOH are different from those from the Uganda Bureau of Statistics (UBOS)’ (MDA-5). While some sectors such as health had elaborated information systems from the village level, other sectors like agriculture lacked such.23 The capacity to design and implement complex multilevel data systems was also considered suboptimal in government. Fragmented data systems were also exacerbated by donor funding practices and reporting requirements. In the words of a non-MOH official:

We (government) do not have an integrated information management system, and it’s a big problem. There are so many scattered information systems, and one of the reasons is the funding. There is nobody to harmonise the funding and say that they want to integrate all the systems. (MDA-5)

Fourth, frequent staff transfers within the civil service were reported to sometimes take away from the collaborating MDAs, institutional memory and allies supportive of multisectoral initiatives.

Government as a learning organisation

Some respondents emphasised that limited evidence generation and use in government undermined shared appreciation of problems and solutions. These respondents concluded that the little efforts to evaluate government policies and programmes undermined evidence-informed decision-making and rendered government, not a ‘learning organisation’. An NSA remarked that:

You are not doing evaluation. So how will you know which policies are causing you the problems? Sometimes, we don’t even know the source of the coordination problem because we do not consciously audit our policies, review them, and keep removing inconsistencies, contradictions, and stuff like that. (NSA-1)

Discussion

Factors that influence intragovernmental coordination

This paper sought to analyse the various factors influencing intragovernmental coordination in Uganda. The study found that significant coordination problems arise from dynamic and complex interactions among different factors. Interdependencies, coordination costs, malaligned interests and institutional and ideational aspects were key factors with mixed influences contingent on the prevailing conditions. The traditional bureaucratic logic of specialisation and associated professional bureaucracy enhanced by NPM principles created an (inter)organisational context that encouraged fragmentation in government.65 The legal-institutional context, especially regarding budgeting, perpetuated siloed planning and programme implementation. Informal institutions such as corruption created a de facto institutional culture antagonistic to coordination. Resource dependence is a major driver of intragovernmental coordination12 15 but was undermined by the desire to control other parties’ resources instead of pursuing mutually beneficial outcomes. Inside government, various MDAs often compete for resources. These realities induce civil servants’ resentment toward collaborative initiatives. Similar to other researchers, the factors above exhibited internal and external dimensions corresponding to the internal or external government context, respectively.1 4 66

Coordination-related transaction costs

The TCE perspective recognises that coordination involves costs inherent in the exchange processes. The MDAs are motivated to minimise costs, which explains why some transactions are internally absorbed by an agency and others are not. This study distinguishes transaction costs between coordinating entities from those incurred by central entities with coordination roles. The costs involved in searching for partners to coordinate with within the government are usually reduced by a clear assignment of mandates. But this is rarely the case.4 67 As our study reveals, there are costs related to building a shared vision, bringing together organisations with various mandates and cross-organisational learning. Consistent with PE perspectives and RDT, the findings confirm that these costs could be positively mediated by trust, considerations of what interdependence entails, a communication strategy and earmarked budgets for coordination.12

The political nature of coordination

The study revealed several manifestations of power dynamics, including competition and conflicts over resources (especially budgets), donor inducing compliance through conditioning coordination on aid, contestations over mandates, framing issues in self-interested ways and adopting behaviours (such as non-compliance to rules) to undermine the power of coordinating agencies. Consistent with the PE literature,60 68 informal norms of demanding teas and transport allowances and instances of corruption, such as misappropriation of funds, were documented.

This study confirms that perspectives, interests and power of coordinating parties matter.60 Organisational and individual interests were reportedly pursued through emphasising professional superiority, lobbying for budgets and external resources, and justifying resistance to change because of favourable power structures. Power and politics underpinned by structural positions in the government bureaucracy were often counterbalanced by the agential power of individual MDA derived from respective legal mandates and control over resources. The incongruence of interests and conflicting incentives among political leaders and technical experts negatively affected coordination by the politicians renegading on positions agreed on by their technocrats. Such dynamics should be managed judiciously.

Ideational factors such as what constitutes interdependences, their meaning and corresponding values, and what are considered to be coordination problems and solutions underlie the political dynamics of coordination.14 Some actors erroneously reduced coordination to meetings and refreshments, negatively impacting commitments to and resourcing of coordination efforts. How one communicates the vision on why coordination matters and how this vision is perceived and potentially shared or not by different MDAs with different mandates, roles, responsibilities and thus interests are critical. Limited information exchange and unfavourable perspectives such as considering coordination to be a purview of agencies higher in bureaucratic hierarchy undermine coordination efforts. Furthermore, fragmented information systems perpetuated by interests and practices of intragovernmental and external actors enhanced information asymmetries between government actors. These conditions (further) hamper effective internal government coordination by increasing transaction costs and distrust arising from opportunistic behaviours.69

New public management, agencification and organisational specialisation

Motivations and barriers to coordination can be understood within the spectrum of history.70 This study confirms observations from high-income settings71 and the rest of Sub-Saharan Africa72 that government architecture is path dependent on the NPM and related public sector reforms of the late 20th century.3 65 73 These findings concur with Cejudo and Michel74 that due to NPM, ‘over time, different policy domains (sectors) developed their own segmented conception of policy problems, appropriate solutions, ideologies and interests’. These realities further underscore the role of institutional and ideational change in advancing intragovernmental coordination.12 The NPM reforms were a constellation of principles and institutional arrangements that emphasised organisational specialisation, narrowing the purview of government and increased privatisation and decentralisation.65 72 This study revealed that the influence of the NPM reforms is still ongoing. Therefore, their mixed (facilitative and constraining) effects on Uganda’s context and motivations for intragovernmental coordination should be anticipated and appropriately managed.65 73 In a broader sense, NPM might have led to flexible decision-making and more heterarchical structures as it diffuses state power by engaging non-state players in governance.75 Those could be genuine and desirable outcomes. However, such goals are achieved at the expense of undermining internal coordination within government due to increased fragmentation and agencification. At the minimum, as the NPM benefits are being harnessed, the detrimental effects on coordination should be paid attention to. Unsurprisingly, counter-reforms toward more (re)centralisation have been adopted in several countries.65 76–78

Donor dependency and extra-government influences

Similar to other studies,48 79–82 the influence of the extra-government context and actors was prominent and realised through shaping development agendas, norms, ideas and resource flows. By illuminating the donors’ power in shaping the development agenda over the last three decades, this study underscores a context particular to donor-dependent LMICs, and a period when certain donors pushed for institutional strengthening of MDAs as a way to bolster (‘good’) governance.65 Donors are strong players and can facilitate or constrain coordinated behaviours by, for example, pushing separate agencies, fragmented funding and data systems that perpetuate sectoral silos.83

Applying a multitheoretical approach: reflections and implications

This paper contributes to the literature on theory building and, more specifically, applying a multitheoretical approach to research84 85 and examines the empirical relevance of the multitheoretical framework proposed by Ssennyonjo et al.12 This work contributes to studies that have used theories to analyse data and interpret findings.33 36 37 86 We demonstrated how a multitheoretical approach is particularly pertinent to intragovernmental coordination. The TCE with emphasis on transaction costs, the agency theory with focus on principal–agent relationships in government, the RDT with its emphasis on interdependence and the PE theory with its attention to politics and contestation over ideas, interests and resources, each provide a partial explanation of the reasons for coordination or not. Taken together, they provide a set of explanations that is consistent with rational choice and power theories.12 This multitheoretical framework underscores that organisations are faced with multiple pressures and interests at any one time. As a result, the coordination decisions rarely come down to a single factor and often emerges from consideration of numerous related concerns, usually leading to compromise and trade offs.81 87 For instance, desirable transitions from the existing norms, practices and structures could be abandoned because of the transaction costs related to the design, implementation and monitoring of new institutional arrangements.

The theoretical perspectives also provide new ways of examining coordination beyond simplistic conceptual frameworks. For example, the usual approach of dividing coordination context into legal, political, economic and social dimensions exemplifies a thematical categorisation for descriptive purposes, an entirely different thing from a theoretical approach.88 A (multi)theoretical approach permits a more profound and multifaceted analysis of complex phenomena. For example, the TCE considers the intragovernmental coordination a constellation of exchanges between the MDAs within and across levels and policy areas (each with interests above and beyond their institutional mandate).50 Agency theory spotlights the inherent interorganisational interactions in government as principal–agent relationships characterised by opportunism, bounded rationality and information asymmetry that are often exploited by actors to hide information (adverse selection) or hide actions (moral hazard).49 50 The PE perspective underscores the internal central government context as a (political) arena, facilitator, constraint and outcome of multisectoral action.4

This study thus reinforces HPSR and especially research on MSA for health that has not adequately drawn on theory.14 89 90 However, there remains scope to consider other implications of the multitheoretical framework in closing the research–practice gap. More so, this study contributes to theory building from LMICs. Several scholars argue that theory building does not originate from empirical case research from LMICs30; the theories tested are from the global north—yet there might be different theories emerging from south case research that are also useful for the global north.14

The study of MSA coordination can be reinforced by applying the multitheoretical framework across diverse organisational environments and country contexts or introducing another ‘rival’ set of theories, for example, critical interpretative theory.91 Using it as a starter programme theory or theory of change, the framework could be applied to study evolutions of coordination instruments, collaborative government initiatives or partnerships involving government and non-government entities.92

Limitations

We note several limitations. First, the study scope focused on coordination at the central government level focusing on the executive activity and broader bureaucracy. The study did not consider the coordination with or within the legislature and judiciary. Second, the study is concerned with coordination within a single tier of government, yet considering the interdependencies and vertical coordination between the central government and subnational levels is critical, especially in the case of devolution.3 Third, the case study of Uganda exhibits contextual differences from other LMICs. LMICs have different political and institutional trajectories. However, applying a multitheoretical framework generates broader relevance of the study findings. It can be used in different case studies of other LMICs and contribute to theory building on governance for health in LMICs, which is undoubtedly needed. Fourth, our choice of theories is not exhaustive. We might have missed certain factors highlighted by theories such as systems theory, complexity theory or critical interpretative theories used in related work.4

Conclusions

This case study applying three organisational theories and a PE perspective to analyse factors that shape multisectoral coordination at the central level (government) in Uganda presents several insights. First, the multitheoretical approach holds the promise of a more holistic approach to exploring factors shaping coordination of intragovernmental efforts and MSA in health. The assumptions and propositions from the four theoretical perspectives provide a broader lens to understanding coordination obstacles and opportunities by emphasising transaction costs (TCE), agency problems and costs (agency theory), interdependencies between stakeholders (RDT), politics, power, interests and institutional and ideational factors (PE perspective). Second, the study further disclosed that these factors may emerge internally within the government or outside from the external context, are contingent and have mixed (positive or negative) influences.

Against that background, the study provides a broad scope of possible actions to enhance MSA for health. For policy-makers, considering theoretical perspectives underpinned by organisational theory and political dynamics offers the opportunity to apply multifaceted strategies beyond the ‘apolitical’ technical solutions. Institutional, operational and process factors interact and evolve dynamically. This approach cautions researchers and policy-makers against normative assumptions that what works at one phase of a collaborative endeavour holds for another. Achieving intragovernmental coordination goes beyond ‘teas’ and meetings. It requires that more time and resources are devoted to guiding the software aspects of institutional change—articulating a common vision on coordination across government. Shaping incentives to align interests, managing coordination costs and navigating historical-institutional contexts are critical. Countervailing political actions and power dynamics should be judiciously navigated.