Introduction

Stewardship and governance, like “resilience” or “strategic,” are “power words” (Frederickson 2005). They sound desirable, are difficult to argue with, and give an automatic advantage in most arguments to the people who invoke them. As a result, both have been stretched by academics, governments, international organizations, consultants, and other ideological entrepreneurs who want the power that comes with its invocation.

This chapter will first separate out stewardship and governance, providing key definitions and making the point that while they might be in the hands of political rivals, they are not intellectually rivalrous concepts. It then presents the results of our review of concepts, presenting the five attributes of governance (which are also among many desirable objectives of stewardship) that emerged as mutually exclusive and able to cover the many activities and ideas classified as “governance.”

Definitions: Into the Mire

Governance has several kinds of meaning. On one hand, it has spread across multiple fields that use it in different ways to discuss topics as different as the proper constitution of a company board and the nature of public management in the Internet age. On the other hand, it is used for a variety of normative, empirical, and mixed projects.

While there have been sporadic uses of the word for many years, it became a common modern concept first in the discussion of management, specifically corporate governance, the organization of power within commercial firms. In the 1980s, it started to pick up a second usage; it was used in political economy research to discuss arrangements in which organizations such as unions, professions, and government collectively coordinated activity (e.g., Campbell and Lindberg 1991). In the aftermath of the Cold War, more academics became interested in it as a descriptive term for systems that produced collective decisions without having clear centers of hierarchical power (as distinct, in some once-fashionable formulations, from “government”). In this capacity, the term drew on and partially displaced perfectly good older terms such as “networks.” In the hands of these scholars, governance came to mean almost anything that generated order without hierarchy; its meanings in transaction cost economics (Williamson 1996), European studies (Marks et al. 1996), international relations (Rosenau and Czempiel 1992), and public management (Rhodes 1997), for example, differed greatly.

International organizations became particularly interested as part of the backlash against structural adjustment lending and, in particular, their role in the Asian financial crisis and its aftermath. Fifteen years of increasingly invasive policy conditionality in the service of structural adjustment failed to produce the desired effects in the structurally adjusted countries (Greer 2013; Woods 2006). They turned to good governance as a solution (e.g., World Bank 1992, 1994). The essential logic was simple enough: reforms, especially those imposed through conditional loans, frequently had serious noncompliance problems, faced serious implementation problems, and had the wrong effects. The response was to blame these problems on the governance – the organization, probity, competence, and coordination – of the countries involved and try to improve that as a part of development or financial rescue (Nunnenkamp 1995).

In 2013, all three preoccupations are alive and well: we have governance as a field of management, including corporate governance and clinical governance in health (Walshe and Smith 2011), governance as a sprawling and contested term applied in endless different ways by social scientists in analyzing the world (Kjaer 2004; Bevir 2013), and governance as a normative concept used when policymakers speak about improving, essentially, international public management (Fukuyama 2013).

In each of these incarnations, governance-speak has two essential uses. One is empirical: the description and analysis of what is. One is normative: calls for how it ought to be. Empirically, governance in almost any account is some form of authoritative coordination, which means decisionmaking and implementation. Such analyses tend to try to capture the mechanisms by which authoritative decisions are made, analyzing the powers, responsibilities, and coordination of professions, insurers, providers, governments at different levels, and the other actors who make and implement decisions in health systems.

Normatively, governance can be termed good, or better or worse, and the parallel normative, policy-oriented literature seeks to improve it by promoting, essentially, various forms of “good governance.” In general, this normative literature is focused on policy interventions and institutional changes. The real solution to corruption, social science makes quite clear, is reducing inequality in society by expanding social rights and economic redistribution (Uslaner 2008; Rothstein 2011). That seems to be beyond the scope of most governance advice, which focuses on the level of individuals (hiring the right people) and organizations, and perhaps legal frameworks (Sabet 2012, 21 for the distinction). Many accounts, of course, mix normative and empirical in more or less coherent, articulated, and useful ways.

Stewardship, by contrast to governance, is a word with a more limited history in health policy. While the word is as old as the concept of a steward – a person entrusted with looking after something – its grand entrance into the global health policy vocabulary came in the 2000 World Health Report (World Health Organization 2000) (WHR), which defined it as one of four key functions of health systems alongside resource generation, financing, and service delivery. The WHR defines stewardship as “the careful and responsible management of the well-being of the population,” and “... the very essence of good government” (Travis et al. 2003 for a lucid discussion in the WHO context).

Separating governance and stewardship is conceptually easier than it might look. Firstly, governance is a structure or pattern, whereas stewardship is an activity. As a result, pursuing an item such as capacity or development or transparency from a long list of policies can be seen as good stewardship or establishment of better governance. A person occupying a position in a system of governance can be a better or worse steward. Secondarily, stewardship is almost always normative in health policy discussions. Governance in the sense of authoritative coordination exists in almost any functional society (by definition), even if it is not good. Stewardship, by incorporating care, responsibility, good government, and the well-being of the population, makes itself a normative rather than empirical concept. Thirdly, stewardship was a concept largely confined to global health policy discussions, while governance is, for better or for worse, discussed in many fields of human activity.

Comparing and Measuring Governance

Measuring the quality of governance has been a preoccupation of scholars and international organizations for some years now, and the result has been a variety of initiatives that attempt to define governance in quantitative, comparable terms. Given that the latest initiatives are the most ambitious yet, the next years should be fertile ones for the quantitative, comparative study of governance.

The largest project is based at the University of Gothenburg. The “Quality of Government” project, as it is known, aggregates a wide variety of databases (its key findings are in Rothstein 2011). A variety of other projects, including the Varieties of Government project based at the University of Notre Dame (Coppedge et al. 2012), try to enhance our comparative understanding and measurement of a wide spectrum of governance indicators (Fukuyama 2013 for a review). These databases, which face the data and coding problems of all large-scale international quantitative comparative research efforts, are mostly focused on general regime types and put less focus on the actual management of health systems.

The measurement of health systems governance is somewhat less developed, since it is not unintelligent to focus instead on actual health outcomes (an imperfect enough set of outcomes) (Smith et al. 2008). The comparison of health systems and their governance is, by contrast, rather more developed. The European Observatory on Health Systems and Policies has significantly advanced comparative health systems research by producing books, written to templates, on the health systems of every country in the WHO European region and a variety of others. Its Health Systems and Policy Monitor is a regularly updated source of information on health policies, from which much can be learned about governance.

Good Enough, or Better, Governance

Two words, three broad traditions of their use, a plethora of international comparative enterprises, and both normative and empirical applications: this is a dispiriting starting point for a discussion of how the vocabulary of governance and stewardship may be used to understand or improve health systems.

The first problem to address is the confusion created by political analysts of many stripes, ranging from entrepreneurial consultants to entrepreneurial academics, who sought to distinguish governance as a type of organization from government. This approach defined governance in terms of self-organization, networks, and a blend of public, nongovernmental, and private actors, rather than “government,” which connoted hierarchy, legalism, and inflexibility. The essential distinction was spurious and misleading; networks were hardly new forms of political organization, in the West or anywhere else, and the hierarchical authority of states and other big organizations such as corporations remained very powerful and effective (Bevir 2013). Here, following on current usage and the international institutions, governance is a description of overall decisionmaking and implementation rather than an ideal type rendering of a particular form of public administration.

The next problem is with the concept of “good governance.” If governance can be better or worse, then it seems reasonable to seek to identify and generalize practices of good governance, whether it is corporate governance activists trying to generalize good recruitment practices for boards or international financial institutions trying to generalize good governance for the recipients of their funds. Two difficulties arise. The first is revealed by the syllogism: if governance is how decisions are made and implemented, then good governance is good decisionmaking and implementation throughout a whole society. The likelihood that the same things, defined with any level of specificity, will constitute good governance in every society on earth seems limited (Andrews 2013). Excessive concreteness is a besetting problem in advice about good governance.

The third problem is that governance, being a power word (Frederickson 2005) whose invocation strengthens all sorts of arguments and claims, therefore has had a wide range of attributes added to it. These are often self-contradictory or hard to derive from either data or first principles. For example, some international organizations view “conflict prevention” as an important aspect of good governance, and others do not (Barbazza and Tello 2014). Does this mean that the WHO regards conflict as part of good governance? Obviously not. Rather, what it shows is that lists of attributes of good governance have a tendency to be arbitrary and utopian. Defining the aspects of good governance is tantamount to defining the good society, and that is questionable on matters of taste and practicality.

Notably, few if any systems show all the attributes that have been assigned to “good governance,” and many highly functional systems have aspects of poor governance – opacity, corruption, nepotism, clientelism, and other problems occur in many places. Few if any countries vaulted into high-income brackets while enjoying good governance as many define it today (Greer and Jarman 2011; Brewer et al. 1999), and a few practices we associated with bad governance have proved flexible and effective, for example, clientelism can mean disruption and bad administration by political jobbers but also allows reformers to put technically skilled people into important posts (Grindle 2012).

The problem, therefore, is the one noted by Tolstoy: all happy families are the same, but all unhappy families are different. So many things have to go right to produce a happy family that the variation within the category of happy families is limited. Unhappy families have many more degrees of freedom. And it is in the realm of unhappy families that policy scholars and policymakers must operate. The solution lies in the simple concept of “Good enough governance.” Good enough governance is a concept formulated by Merilee Grindle, who pointed out that many lists of governance attributes have an arbitrary and utopian character (Grindle 2004, 2007; Thomas 2015).

Drawing from this, a more intellectually and practically satisfying approach to governance is to view governance not as a desirable end state but rather as an activity that can be carried out in different ways with different effects. This diagnostic approach views governance as a phenomenon that exists in essentially all societies and sometimes causes a problem for something else. Governance problems can be diagnosed as a reason for policy failures, and strengthening one aspect or another of governance can remedy policy failures. Likewise, some policies are just not sustainable in some systems; governance that is good enough for maintaining basic public health functions might not be good enough to operate sophisticated quasi-markets for health care.

In other words, rather than insistently defining good governanceit makes more sense to identify aspects of governance that improve the ability of health systems to achieve a sustainable balance of equity, access, and cost containment. So, then, what are aspects of governance that influence the ability of health systems to achieve their goals, and which can in some cases be improved? Or, on the other side of the coin, what is a governance problem (as distinct from some other kind of problem), and what is a detailed taxonomy of governance problems that might need understanding or remedy?

Attributes of Governance

The first question in using governance analysis to improve policies and systems: is the challenge, or problem, or opportunity one of governance? There are other reasons programs fail. They can be fundamentally bad ideas (though high-capacity, participative, transparent governance might reduce the odds of bad ideas being adopted). They can be underfunded. They can also lack political support.

By a process of elimination, a workable, funded, and supported policy that fails suggests a governance issue. More positively, do problems appear to lie in the decisionmaking and implementation systems of society? If so, that means the problems lie in governance.

More specifically, our review found five key aspects of governance that matter and in many cases can be strengthened. They are not a list of attributes to which every society should aspire; they are, rather, five aspects of health systems that influence the success or failures of policies. One of the remarkable aspects of the governance literature is that, beneath a level of apparent conceptual confusion, the same words and concepts constantly recur. In other words, despite many different terms and many different lists with different inclusions and exclusions, and many different conceptual hierarchies, the same five issues recur. We sorted them into groups with minimal overlap that scholars or policymakers interested in governance should consider (Greer et al. 2016; Greer et al. 2017). The result is the TAPIC framework, for its domains of transparency, accountability, participation, integrity and capacity any of the five might be the first or most important issue, and all can exist relatively independently of each other (accountability without transparency, for example, is the norm in both medical care and automobile repair). The literature review and analysis is presented in (Greer et al. 2016). Case studies exploring and showing the uses of the TAPIC framework can be found in that book, and in (Jarman 2017, Wolfe et al. 2017, Exworthy et al. 2017, Trump 2017, Vasev 2017, Willison 2017 and Greer et al. 2017)

Transparency

Transparency involves two things: making decisions clear and making clear grounds on which decisions were made (Woods 1999). At a minimum, this means the kind of basic publicity long familiar in functional governments – official notifications, open meetings, and latterly informative websites that make policies and policy processes understandable.

There are a variety of problems with such a simple form of transparency, however; for a start, as every consumer knows, “fine print” can look transparent and effectively hide companies’ actions. Transparency can be taken too far; decisionmaking necessarily involves both deals and ambiguity, and problems arise if transparency displaces real decisionmaking into shadows or becomes a weapon for those who want to replace argument and prioritization with some more mechanistic (Best 2005). It also has the problem that policy information can be intricate, and efforts to simplify it can also distort it (as frequently happens with both politics and website redesigns). The result is that simple notification should probably be flanked by devices that permit informed access to the policy process so that informed journalists, NGOs, citizens, and experts can contest decisions and their grounds. These mechanisms can include inspectorates, ombuds procedures, public data releases, and freedom of information laws.

Effective transparency should improve policy by enhancing accountability and participation, deterring or quickly identifying corruption and incompetence, and making policies more predictable. The result, in theory, will be trust that an organization will not be erratic and in constant pressure to be competent.

Accountability

Accountability is a relationship in which an actor (such as a government agency) must account for its actions to a forum (such as a legislature) which can sanction it. In other words, it has three key attributes: actions, reporting, and sanction. A good accountability relationship means that the interests of the forum (legislature, population) is always in the mind of the actor, but the actor has autonomy to formulate superior solutions. It can also allow productive innovation; holding somebody accountable for outcomes within limits rather than process can produce learning and better policy outcomes in general (Sabel 2001; Behn 2001).

Mechanisms that policymakers use to achieve accountability are diverse, including contracts; reporting requirements; financial mechanisms such as pay for performance; laws that specify objectives, reporting, and mechanism; competitive bidding; organizational separation such as purchaser/provider splits; conflict of interest policies; ombuds processes; legislative oversight and committees of oversight, and regulation including the establishment of dedicated regulatory agencies. Each of these focuses on increasing the extent of reporting and the ability of the forum to sanction the actor.

Accountability is not the same thing as a principal-agent relationship, which favored form of economic modeling. In a public sector principal-agent relationship, a principal chooses an agent to carry out its wishes (Smith et al. 1997; Besley and Coate 2003). Governance, in this analysis, is better insofar as it shortens and clarifies principal-agent relationships. There are two key problems with this style of analysis. The first is that frequently the relationship is hard to characterize in that way – it might actually be a fiduciary model rather than an agency relationship. The second is that it is essentially normative rather than political; it assumes that there should be a clear principal, agent, and instructions. A quick reflection on, for example, the many missions of a hospital shows the empirical limits (Marmor 2001).

Participation

Participation means that affected parties have access to decisionmaking and power so that they acquire a meaningful stake in the work of an institutions (Woods 1999). Participation has many normatively desirable aspects – it is the basis of democracy, after all – but there is also a pragmatic case for participation of affected parties in decisions that spans political regimes. That is simple: participation helps to reduce or avoid the problems that emerge when key affected groups resist a policy or when a policy is made without knowing what they know. For example, complex medical payment incentive systems do not work as intended if they are made without understanding how doctors work and are paid (a common problem in “pay for performance” schemes). In the worst case, it makes it clear what depth of opposition a policy will face once enacted.

There are a variety of well-established participation mechanisms, as well as a very large and notably confused literature on public participation in health that rarely explains the point of participation (for a critical discussion Stewart 2013) and some experiments in novel forms of public participation, such as participatory budgeting, whose popularity outside their places of origin is clearer than their effectiveness (Seekings 2013). Established mechanisms of participation include stakeholder forums, public consultations, elections, appointed community representatives on boards, and legal remedies (e.g., legislation that allows aggrieved outsiders to litigate processes). They can also include research, e.g., surveys of local opinion about a given option. When affected bodies are other governments or organizations, advisory committees, partnerships, joint budgets, and special forums for consultation are effective mechanisms for ensuring that different governments will be aware of decisions and make their views clear.

The benefit of participation is the potential creation of “ownership,” i.e., a sense among affected parties that they have a stake in the success of an initiative. Without ownership, there is a real risk of sabotage, lassitude, or simple ignorance, all of which amount to implementation failure. There is also the potential benefit of increased legitimacy – the sense that decisions are taken in ways that reflected the relevant interests.

Integrity

Integrity is one of many words for the key attributes of a well-run modern bureaucracy: processes of representation, decisionmaking, and enforcement should be clearly specified; all members should be able to understand and predict the processes by which an institution will take decisions and apply them; and individuals should have clear roles and responsibilities. In other words, an organization with a high level of integrity is meritocratic, separates the person and the office, and is not corrupt. These are the bases for well-functioning, long-lasting trustworthy organizations.

Mechanisms policymakers can use to promote or entrench organizational integrity include internal audit (so that money moves as intended and can be traced), clear personnel policies (regular hiring, job descriptions, and procedures to weed out flawed people), a clear mandate for each organization, a clear and reliable budgeting process, administrative procedures such as document management and minuted meetings, external audit (to put a check on people within the organization), and a clear sense of organizational roles and purposes. Many of these policies, if added together, are bureaucracy – for better or for worse. The challenge of public management is to gain the benefits of bureaucracy in terms of merit, impartiality, and efficiency without risking too much wasted effort or incompetence.

Policy Capacity

Finally, most accounts of effective health governance include a discussion of policy capacity: the ability to develop policy that is aligned with resources in pursuit of societal goals. Policy capacity is a property of what Edward Page calls the “policy bureaucracy,” that part of an organization, especially a government, whose purpose is to produce policy (Page and Jenkins 2005). Just as a health policy initiative can run into trouble for a lack of medical staff, it can run into trouble for a lack of policy staff who are capable of identifying, synthesizing, and analyzing a wide variety of information in order to spot problems, make the case against ill-considered policies, and work through the procedural and practical challenges of implementation. It can look good to reduce policy capacity – civil servants at the heart of the state do not always have public sympathy – but it can have negative consequences in the form of poorly thought-out policies.

The development and improvement of policy capacity is a central preoccupation of public management scholarship, and the list of tools for doing it is long. It includes mechanisms to produce intelligence on developments in the system and its performance, so that policymakers can identify and react to problems and intelligence on process such as budgetary and legal issues (all too often neglected in health policy analysis), research and analysis capacity (trained staff who can conduct or commission research and deal with literature and outside experts), staff training (e.g., so that a doctor hired into a health ministry can learn about budgeting and law), strong hiring procedures that balance merit and responsiveness in the central policy bureaucracy, procedures to incorporate experts with their different career structures and incentives, and, all too often forgotten, extensive capacity for purchasing and managing relationship with outsiders such as regulated industries or government contractors. This long list suggests something important: while policy bureaucracies are routinely dwarfed by the systems they manage and they go beyond the minister’s immediate office. Civil servants further from the minister, and from the glamor of politics, fulfill an important role and can respond to investment and organizational development.

A Diagnostic Approach

Reading scholarly and grey literature, almost everything framed as a component of good governance or as an attribute of governance in general, can be fitted into these five categories. If we use them as a diagnostic tool (before or after a problem arises), then we can first see if a policy failure, or risk, depends on decisionmaking and implementation and then work out what kind of governance issue exists and might be remedied – if, for example, the problem is of sabotage and poor implementation by excluded interested parties, then greater transparency and participation might be called for. It is less productive to elevate them, or any other framework, into good governance, for the simple reasons that there are tensions between them, all of them can be taken to extremes (e.g., transparency can make productive dealmaking impossible), and not all of them will mean the same thing or have the same salience in every system (e.g., integrity is much less of an issue in Northern Europe than in most of the rest of the world). We can, however, try to use the TAPIC framework for diagnoses not just of specific policy problems but of policymaking problems. This should in turn reduce the likelihood of unworkable policies being adopted, or workable policies adopted without adequate finance.

Conclusion

Governance and stewardship might seem like hopelessly fuzzy concepts, but the exercise of grouping the many things said about them reveals five relatively coherent attributes of a health system that are the object of policies for improvement and that can have an effect on the ultimate cost, quality, and access of health.