The TCE theory

The TCE posits that the actors choose their position to coordinate or not and organise interdependencies on considering costs of exchange relations.31–33 ‘Transaction’ is the main concept in TCE. The costs associated with these transactions become a central motivation of why and how to govern those exchange relations. TCE explains weak coordination between health and finance ministries by referring to high transaction costs (TCs). For example, physicians and economists use different language, care about different things and have trouble identifying shared values.34 These high TCs are a major obstacle to coordinated intragovernmental action. Coordination within government often gets down to exchanges between different MDAs. These exchanges exist within the following relations (1) politicians and bureaucratic officials, (2) one sector and another, or (3) government and non-state actors. Examples of exchanges in multisectoral arrangements include information, financial resources, staff and mandates.35 Understanding these exchanges becomes an entry to explore the obstacles and opportunities to enhance coordination among government entities.36

MSA for health is premised on the fact that some actions that contribute to health and well-being are better suited for organisations outside the health sector.34 This, in practice, might mean referral of duties from the health sector as a means of reducing workload for the sector. It also mandates the Ministry of Health (MOH) and government to develop mechanisms to coordinate the exchanges for MSA for health. TCE assumes public organisations to decide whether to coordinate (buy the goods or services from others) or not (to absorb the production function internally) based on perceived costs of exchanges.37 The TCE argues that organisations balance both internal production costs and TCs to keep these costs a minimum.30

Additionally, coordination will develop after internal considerations to improve efficiency and not desire to conform to mandate requirements or propensity to control or pursue mutually beneficial goals.38 Creating parallel or duplicate structures in the MOH to address cross-cutting issues is analogous to the ‘make’ decisions. Relinquishing some responsibility to other MDAs is analogous to a ‘buy’ decision. Therefore, the cost of the above actions by the health sector (MOH) is critical for decision-makers.

Organisations often perceive addressing cross-cutting objectives as resource-intensive39 and prioritise the mechanisms that minimise costs. The TCE predicts that the choice of coordination mechanisms in government dwells largely on the need to minimise the cost of coordinating transaction exchanges.27 31 40 The TCE logic presents the market (outsourcing) and firm (internal production) as two extremes of coordination mechanisms. TCE also argues that the presence of bounded rationality reinforces the decision not to coordinate. The more specific the assets involved in the exchange relationship, the high uncertainty and bounded nature of the actors’ rationality forces actors to prefer internal production of goods and services to minimise the cost of coordinating transaction exchanges.31 Thus, from the TCE perspective, hierarchical coordination arrangements within the entire government or a selected ministry are motivated by the need to reduce their business costs.40

Market exchanges could arise from referrals and outsourcing efforts.30 The frequency of exchange transactions and measurement challenges also matter in the choice of coordination mechanisms. When exchanges are frequent and measurement of results is difficult, organisations will prefer internal hierarchical-based coordination instruments. At the same time, infrequent transactions favour market-based coordination mechanisms.41 Hierarchical tools are a mainstay feature of coordination of MSA. Recent examples include the centralised incident command systems activated in countries like Uganda to coordinate the national COVID-19 responses.42

Principal–agent or agency theory

PAT looks at coordination arrangements as constitutive of principal–agency relationships. The principal–agency relations exist where a principal engages an agent to undertake certain tasks because the former have limited capabilities and are less efficient performing them.41 43 PAT draws attention to the several principal–agent relations within an intragovernmental set-up. For instance, politician and civil servants are comparable with principals and agents, respectively. Politicians often articulate their broader goals and leave public officials to implement. Hence, this delegation causes uncertainty of results and the principal’s (politicians) tendency to promote mandated multisectoral coordination to govern agents’ (MDAs) actions.39 Other examples of principal–agent relations in the context of MSA for health include those (1) between the health sector (specifically MOH) and MDAs in other non-health sectors and (2) a relationship between a central level agency (like the Ministry of Finance) and the MOH.

Agency theory posits that both the MOH or other MDAs (either as the principals or agents) are opportunistic and aim at maximising their interests. Examples of opportunistic behaviours include limited efforts by agents (moral hazard) or misrepresentation of abilities and hidden information (adverse selection). Empirical evidence suggests that agents usually have no incentive to coordinate and hence prioritise ‘tasks which are of primordial importance to them, for which they are accountable and benefits them the most’ (39; p 241). In addition to the coordination costs, the agents calculate the potential effects on their power and resources to decide whether to collaborate or not.39 The principal has to perform close monitoring of the agent because disparities in information lead to agents’ tendency to engage in opportunistic behaviour.44 Information systems become a default tool for advancing coordinated actions. The principal can also achieve coordination by creating incentives to induce cooperative behaviours among the agents. The agency theory underscores contracts as coordination tools available to principals to align the agents’ actions to their objectives. Two types of contracts are defined based on the ability to measure outcomes and the extent of the risk in getting results, namely (a) behaviour-based contracts or (b) outcome-based contracts. Behaviour-based contracts are preferred in instances of difficulty to measure outcomes, task programmability (ability to specify agent action in advance), uncertainty about the outcome, long-term relationships and risk-averse agents or principals. The principals monitor the agents’ behaviours, such as participation in joint meetings and procedures as a proxy to the commitment to work in a coordinated manner.

Examples of behavioural-based contracts include coordination instruments for collective actions from health policy and systems research, including (a)collaborative planning through tools like health technology assessments45, (b) joint implementation arrangements such as one-stop shops, and (c) undertaking joint monitoring and reviews.46 Policy development stages such as agenda setting, policy formulation, implementation and evaluation can be useful entry points to explore the coordination process.47 A recent paper by Okeyo et al48 looking at coordination in the adoption phase of an intersectoral First 1000 Days Initiative targeting early childhood in South Africa is an example. Health policy and systems research (HPSR) literature on Health in All Policies (HiAP) approach covers a broad array of coordination issues during the adoption and implementation of policies.18 28 49 Examples of policy or programme integration efforts include adopting national multisectoral nutrition policies,7 anti-tobacco legislation,50 and school health policies and programmes.51 Another common practice is monitoring the participation in the provision of social and health services around user groups such as children, elderly and women, such as multisectoral maternal and child health programmes.13

Outcome-based contracts are preferred where measuring outcomes is easy. They are also applicable when there is more certainty about outcomes, non-risk-averse agents, high goal conflict, limited task programmability and short-term relationship.27 Outcome-based contracts use performance-based contracts and incentives to reward outcomes. They are consistent with market-based approaches such as result-based management tools advanced through new public management.52

The main limitation of the rational choice perspectives is their unidimensional conception of human behaviour. They assume an actor’s actions to be static and predictable. These theories take for given the institutional context and the social and political process through which IORs occur.27 30 From the above account, rational-based theories must be complemented with theories that capture the dynamic interaction between agency (actors) and structure (context) and the values and power elements that characterise interorganisational relationships. The next section covers the RDT, followed by the PE perspective.

Resource dependence theory

The RDT postulates that organisations seek to find coordination mechanisms to ensure a smooth and predictable flow of resources under other entities’ control. Coordination is driven by the need to control external resources. The MOH and other MDAs consider mutually beneficial collaboration based on the interdependence that arises because each MDA possesses resources needed by another.53 The external environment includes, among other elements, the various government and non-government organisations that constitute the interorganisational context. To survive, organisations should continually seek to influence this environment to ensure stability and predictability of the flow of the resources outside their control. Coordination efforts ‘serve as coping strategies to forestall, forecast, or absorb uncertainty to achieve an orderly, reliable pattern of resource flows and exchanges’ (38; p 245). Consistent with this driver for coordination are coordination mechanisms such as stakeholder dialogues that serve as platforms to support information and resource-sharing and joint action.54

The RDT underscores the prominence of interdependence among different sectors. Mutual reliance on achieving common and specific development goals becomes a launchpad for effective coordination of MSA. For instance, objectives of non-health sectors such as education, agriculture and economy are dependent on having healthy and productive populations.24 From the RDT perspective, the expected coordination mechanisms are network-based, voluntary and based on negotiation, not coercion or imposition. However, higher authority can shape networks by determining their composition and processes. Thus, coordination in networks might be imposed, induced or mandated by a higher authority.19 The interdependence that arises when each actor possesses resources needed by others lead to IORs by the contingence of reciprocity. Resources involved in MSA for health as per RDT are broadly considered. They include finances, human resources, information, social support, authority, technology, access to political offices and acknowledgement.30 Reciprocity underlies the pursuit of interests and goals that are mutually beneficial. Coordination between the MOH and other MDAs is emphasised instead of control, power or domination in asymmetrical motivations. IORs based on the reciprocity contingency assume that resource scarcity leads to cooperation instead of competition. The process is characterised by harmony and mutual support rather than coercion and conflict.

Network theory and related concepts such as network governance, network management or networked government accentuate the nested nature of public sector organisations.55 56 Like RDT, network theory also acknowledges the dynamic interactions between (potential) collaborators and their environments. Consistent with Berg,57 power-based theories hold that coordinated action occurs in two sets of environments, namely ‘internal’ and ‘external’ organisational contexts. The two settings correspond to the two categories of coordination challenges faced in practice-vertical and horizontal difficulties. Vertical challenges arise because each coordinating entity is embedded in a web of other organisations. Through stakeholder analysis,58 the type and network of actors needed to address a problem or actualise the solutions can be identified and presented in a matrix linking issues and actors.36 Researchers can use tools such as social network analysis59 60 to describe the intensity and scope of the IORs. The WHO EURO governance mapping tool is another example of such instruments.61

In the context of government, each MDA or part thereof is part of larger systems and networks. Horizontal challenges arise because there are variations among organisations at the same hierarchical level regarding goals, strategies, language and work routines. For example, how results are registered and reported may vary across organisations.57 These internal and external linkages shape the action and decision spaces within which the actors can operate.

Asymmetrical motives in forming interorganisational interactions as evidenced by the desire to exercise power or reluctance to give up control over other organisations’ resources have been documented (38; p 245). Power and control explain the inclination towards hierarchical coordination arrangements. The public sector coordination should be seen to operate within a political (contested and negotiated) environment characterised by manipulation, conflict, information distortion, coercion and exploitation.62 Empirical evidence further shows it is difficult to achieve coordination in networked relationships. Documented issues related to which sector should lead or where the secretariat for MSA should be housed.63 Therefore, considerable time and energy are required to establish and maintain overtime mechanisms to facilitate coordinated actions to address interdependent development problems.5 24 64

The PE perspective

Although some of its elements have been discussed under the preceding theories, the PE viewpoint is increasingly considered a separate analytical lens in many health policy and systems studies24 and is worth examining in depth. The PE perspective considers coordination a political phenomenon marked by conflict, contestation and negotiation over ideas, interests and power. IORs evolve because of political (ie, negotiated and often contested) processes. Various organisations are competing to shape policies, institutions and resources for coordination.5 24 65 The PE lens provides a foundation to study how power embedded in institutions and ideas is exercised by actors to shape the how and why of multisectoral coordination.63

The institutional–structural context embodies power and determines what agents (individual or collective actors) can and cannot do and consider as achievable by providing resources and constraints. Actors exercise agency to manoeuvre through a complex institutional and ideational context determined by power dynamics. This leads to conflicting interests and policy actions. One illustrative example of incoherence in priorities across government ministries concerns health workforce planning in Uganda.66 Whereas the MOH advocates for the recruitment of additional health workers, the Ministries of Finance and Public Service prioritise cost containment through establishing a wage bill ceiling. Conversely, the Ministry of Foreign Affairs advances policies for labour export.

Consistent with power-based theories, asymmetrical motives in the formation of IORs are characterised by the desire to exercise power or reluctance to give up control over other organisations’ resources.27 Government MDAs tend to resist external control by another agency and resort to turf wars over mandates to maintain independence, thereby undermining coordination efforts.5 For instance, Mauti et al documented that the HiAP strategy in Kenya was hampered by concerns by non-health sectors that such initiatives perpetuate asymmetrical intentions of ‘health imperialism’.67

PE supports generating politically feasible solutions and ensures realistic expectations of what can be achieved.62 Coordination must be proactively sought, and collective efforts require political work to be sustained. The PE perspective illuminates the political actions that MDAs engage in to facilitate or constrain coordination among government entities. Examples of political actions include dialogue, coalition building, backdoor deals, bureaucratic politics, lobbying, campaigning, turf wars and framing.65 By implication, a coordinated approach can be strengthened using different forms of power. Regulative institutional elements explain necessity as the main contingency for formation coordination arrangements and mechanisms.68 In this case, coordination is mandated by higher authorities (higher government agency, regulatory body, professional body), and organisations respond positively to meet legal or regulatory imperatives. The actors consider the anticipated consequences of non-compliance (eg, resource loss) to conform to mandated relations.

Power entrenched in ideas (ideational power) is critical for coordination. Developing a common mental frame (vision) is vital for coordination to ensue. Some efforts on coordination have focused on the strategic framing of problems and solutions.69 70 Institutionalisation and entrenchment of such frames are associated with the development of broader government reforms in developed countries such as ‘Joined up’ government in the UK, ‘Whole of Government’ in Australia and Horizontal government in Canada.19

Normative elements of institutions motivate actor (in)action based on the notion of appropriateness.68 Premised on this logic, organisations will coordinate to improve their legitimacy by conforming to prevailing institutional environments (norms, beliefs, values and expectations) from external actors. Engagement in coordination can be considered an organisation’s motive to ‘demonstrate or improve its reputation, image, prestige, or congruence with prevailing norms in its institutional environment’ (38; p 246). Societal values such as trust have more explanatory power than other reasons for coordination decisions. For instance, trust can be a particularly important mediator of TCs. Where trust exists, the need for monitoring performance and information is lessened.30 These institutional elements derive conformity and lead to more voluntary coordination arrangements.

In summary, this section covered four theories useful to understand and explain obstacles and opportunities to intragovernmental coordination and MSA for health. The four theoretical lenses implicitly or explicitly recognise that government organisations exist within an environment constituent of other organisations, although emphasising different elements. Several factors, external and internal to government, interact in complex ways to shape coordination within the government in LMICs.18 These aspects are worth attending to and are discussed in the next section.