The challenge of policy coordination

Abstract Policy coordination is one of the oldest challenges for governments, but has become even more important as the problems confronting governments change, and the ideas of “New Public Management” are diffused. This paper examines the causes for coordination problems and the mechanisms that may be available for improving coordination. It concludes by discussing the limits on coordination as a solution for the problems of governing.

to fragment further and already fragmented public sector. Therefore, prime ministers such as Tony Blair began to call for a more "joined up" government, or in the extreme a more holistic government (6, Perri et al. 2002) and the government of New Zealand had a program to "restore the center".
The other factor placing pressures for increased coordination was the emergence of difficult problems that could not readily be solved through the actions of any individual public sector organization. These problems, sometimes characterized rather loosely as wicked problems, require substantially greater coordination efforts than do relatively tame problems that fall neatly into the domain of a single government organization. Attempting to deal with a problem such as climate change or sustainable development requires the involvement of much of government, and hence coordination.
This paper will emphasize horizontal coordination within the public sector, but in all political systems, some vertical coordination is also important. This is especially important in federal regimes in which sub-national governments may have substantial autonomy and, therefore, the central government may need to find some means of steering the system as a whole. Vertical coordination also has become very important in the European Union, which represents something like a federal political structure (Bolleyer and B€ orzel 2010).
The remainder of this paper will begin by examining several alternative definitions of coordination, and at several possible ways of measuring coordination. I will then discuss why, when coordination is obviously important in public policy and administrationto the extent that some have referred to it as the "philosopher's stone" that can solve the problems of public administrationwe do not do a better job of coordinating. Further, if we want to coordinate, what instruments are available. I will end by pointing out that although coordination is important, it also has its limits and in some cases can be harmful to the achievement of important policy goals.

What is coordination?
Everyone speaks positively of coordination, but what do we mean by the term? One standard definition from Charles Lindblom (1965, 154) is A set of decisions is coordinated if adjustments have been made in it such that the adverse consequences of any one decision for other decisions in the set are to a degree and in some frequency avoided, reduced, counterbalanced, or outweighed.
In other words, coordination occurs when decisions made in one program or organization consider those made in others and attempt to avoid conflict. That is what Fritz Scharpf (1994) has called negative coordination. Positive coordination, however, would require the organizations to go beyond simply avoiding conflicts and to seek to find ways to cooperate on solutions that can benefit all the organizations involved, and their clients. Coordination problems are not just about conflicts, however, and may also arise because of the redundancy of programs, or the existence of gaps in coverage because programs have not been coordinated. Those problems also can be dealt with positively in order to produce better services.
Although positive coordination can be a major contribution to administration and policy, we can also think about yet another level. Strategic coordination would involve the coordination of programs around broad strategic goals of government. Achieving major goals such as improving the health status of a population or advancing sustainable development will require the involvement of many actors within the public sector. Further, for strategic coordination, the action will be required prospectively rather than retrospectively as is often the case for other forms of coordination.
Another way to address coordination is to consider the levels of coordination that may be achieved. Les Metcalfe (1994 , Table 1) developed a scale of coordination ranging from almost total independence of programs to very close policy integration. This scale was developed to understand coordination within the European Union, but is relevant for almost any government. Some, however, such as those with strong central institutions or with dedicated individuals may be more capable of reaching the higher levels. Very few sets of organizations would reach the top levels of this scale, but it does provide a useful standard against which to compare real-world patterns of interaction among programs and organizations.
Defining coordination also raises issues about whether would-be coordinators should focus their efforts on the top or the bottom levels of the organizations involved. On the one hand, if programs can be harmonized at the top of the organizations involved then the problems may be solved before the programs are implemented. However, attempting to produce coordination and policy integration at that level may produce political conflicts over resources and the interpretation of laws. On the other hand, although there may be difficulties at the implementation stage, producing coordination may be easier when there are real clients with real needs involved. The public servants involved then may have greater pressures to produce solutions to the conflicts and the missing elements in programs.

Why care about coordination?
While coordination is usually thought of as a good thing in itself, there are more practical reasons to invest time and political capital in coordination. These include 1. Duplications: Programs may do the same thing, or may ask citizens for the same information again and again. This duplication can produce unnecessary costs for government, and lost time for citizens and businesses.

2.
Contradictions: Different organizations, often for good political reasons, will implement programs that are directly contradictory. Environmental agencies may want to reduce intensive agriculture while agriculture ministries will want to increase production. 3. Displacement: One organization will make decisions that create problems for other organizations, without consultation. For example, after the 9/11 attacks in the United States, the new Department of Homeland Security said passports were then necessary for Americans to travel of Canada and Mexico. They, however, failed to inform the Department of State that issues passports, and which was faced with the need t o issue millions of new passports. 4. Emphasizing vertical management: The New Public Management has emphasized management within individual organizations. But that has tended to de-emphasize horizontal management involving other organizations. If an organization is being assessed directly on the performance within their organization they are less likely to invest resources in helping others. Therefore, some emphasis on horizontal action is required to augment the emphasis on the vertical. 5. Changing demands: Client groups such as children and the elderly which have demands for a variety of services have become more important politically. Therefore, governments must find ways of providing integrated services to these population groups. 6. Cross cutting problems: As well as the mobilization of client groups that have needs for a range of services, some of the most important problems facing contemporary governments cut-across the usual lines of departmental responsibilities. 7. Simple tidiness: Finally governments may be interested in being better coordinated simply to appear more capable and to build confidence among the public. Given that levels of confidence in government have been declining, it is important to do anything possible to improve the image of the public sector.

Why do not we coordinate?
Many governments confront all or most of the pressures for coordination mentioned above, and might be expected to invest more heavily in coordination, but yet there are numerous examples off the failure to coordinate, and many examples of now even trying very hard to produce more coordination. So why do governments prove to be so resistant to coordinating more extensively. Again, there are a range of reasons for the persistence of the "silos" and "stove pipes" that exist in most governments: 1. Specialization: Perhaps the principal reason there is less coordination is that its antithesisspecializationis also an important value in government. Herbert Simon (1947) argued that most reform in the public sector was moving back and forth between positive values, and specialization versus coordination is one of those dichotomies (Bouckaert et al. 2010). 2. Power: In government and other information processing organizations, information is power so there is insufficient sharing of information. Even if the goals of an organization may be advanced by information exchange, many organizations perform to horde information (Stinchcombe 2000).
3. Performance management: One element of the New Public Managementperformance managementhas had a particularly negative effect on coordination. By setting targets of an organization, it will tend to ignore collective goals. 4. Turf: Organizations want to defend their budgets, personnel and policies, and fear that coordination with other organizations will endanger their "turf". 5. Beliefs and ideologies: Specialized organizations in government will be popular primarily by individuals with a belief in the mission of the organization. That ideology within the organization can be reinforced by professional training and the tendency of professionals to have a particular conception of policy problems and the possible solution to those problems. 6. Politics: Given that we are talking about coordination within the public sector politics will be involved. In coalition governments, if ministries are controlled by different political parties there may coordination problems. Likewise, in federal systems, if different provinces or states are controlled by political parties other than that in control at the national level, there may be coordination problems. 7. Accountability: In the public sector, accountability, like coordination, is a virtue.
But strict financial and legal accountability may make coordination more difficult. If auditors cannot track the money and parliament can not assign responsibility for actions, then coordination may be stymied.

How do we achieve coordination?
Governments have invested a great deal of time and energy in attempting to achieve better coordination, and have a number of mechanisms at their disposal. All of these mechanisms have some virtues, but none is a panacea. Some of these mechanisms for coordination depend upon more or less top-down imposed forms of coordination while others may occur more through individual interactions and bargaining among the actors involved.

Networks
Some coordination can occur through networks, especially networks of career civil servants. These networks need not be formalized but may develop over time through interactions among civil servants who work together and know each other well enough to be able to coordinate outside of official channels. Unfortunately, however, the opening of career public services to outsiders as one part of the New Public Management reforms has to some degree weakened these internal networks. The existence of networks between social actors and government can also serve as a mechanism for coordination. These networks can provide for coordination from the bottom up, with the various social actors involved having information about what the different government organizations with which they interact are doing, and identifying contradictions and lacunae among them. This form of coordination does, of course, depend upon the willingness of the public sector organizations involved to heed the information being given to them by their partners in the networks and to work on better policy coordination.

Collaboration
While most mechanisms available for coordination depend upon structures and the interactions of actors, this alternative depends more on ideas. As noted above, one of the causes of poor coordination is that organizations have different ideas about good policy and the ways in which to address problems. Health organizations will see the problem of illegal drug use very differently than will law enforcement organizations. If these organizations can reach some basic agreement on the nature of the problem and perhaps on the means of addressing the problem, then more effective coordination, and more effective policy, may emerge.
Eugene Bardach (1998) has referred to the process of creating that common understanding of the problem as "collaboration". Somewhat similarly Sch€ on and Rein (1994) have discussed "reframing" as the means of addressing intractable policy problems. In both instances, the logic is that if a common frame for a problem can be identified that multiple actors can accept and work with, then the coordination issue, or perhaps more accurately policy integration issue, can be solved. That common frame is the product of bargaining among the representatives of the various organizations involved.
Resolving coordination problems through collaboration or reframing can be very difficult and time consuming. There are often deeply embedded ideas about policy that must be reconciled across actors. Further, when attempting to find some compromise position among actors the bargaining may produce a solution through the lowest common denominator and produce very little process in actually solving the problem (see Scharpf 1988). But, if this process is successful, it may produce greater coordination than the more structural solutions that are usually applied to the coordination issue.
This means of coordination, like networks mentioned above, may in the end depend upon individual action to produce higher levels of coordination. Individuals have to be willing to bargain over the definitions of problems and programs in the collaborative model. And individuals have to be able to work together in networks if that model is to be successful. Structures for coordination are important, but so too are committed and capable individuals to populate those structures.

Hierarchy
When faced with a coordination problem the usual remedy selected by government involved hierarchical authority coming from the center of government. This is perhaps a natural reaction given that individuals occupying positions in the center of government are most likely to identify the needs for coordination. Further, they will tend not to have commitments to any particular agency or department and, therefore, will not be constrained by loyalties or beliefs in particular ways of addressing policy problems.
The center of government actually has a very large repertoire of instruments to use when confronting coordination problems. Some of these instruments have been in place for decades, while others are relatively recent innovations. All, however, depend to some degree on the authority of actors in the core executive-presidents, prime ministers, and their allies. Therefore, these mechanisms are often as much political as they are administrative. Coordination is about setting and implementing priorities as well as about merely getting organizations to work together smoothly and effectively.
Central agenciesmeaning organizations that supervise and support line agenciesare perhaps the most common mechanism for attempting to create coordination. Organizations such as ministries of finance, budget offices, personnel offices, and the like fit into the category of central agencies (see Dahlstrom et al. 2010). These organizations are charged with making the rest of government pursue, in so far as possible, the priorities of the government of the day, and with using the budget and their influence over legislation to make that happen. Coordination may not be the central responsibility of these organizations, but it is certainly one important task.
Second, there are mechanisms within the government cabinet that can be used to produce greater coordination. Cabinet committees are mechanisms for bringing together ministers representing their departments in order to produce some collective policies (see Hustedt and Tiessen 2006). Also, most parliamentary systems use junior ministers who may be given responsibility for client groups such as women or children. They do not have to defend their ministries as a minister would, but tend not to have many resources to encourage cooperation from the ministers who do. Similarly, there may be ministers without portfolio who have cabinet status but again do not have a ministry to defend in negotiations. These officials can bargain across ministries and attempt to generate cooperation.
Governments can also use structures to attempt to enhance coordination. For example, a number of governments have created "superministries" that bring together a number of related organizations to create internal coordination. Examples would be the Department of Homeland Security in the United States and Employment and Social Development Canada (see Kettl 2007). While it appears logical as a means of producing coordination, the components of the large organizations may retain their own internal cultures so the coordination problems may be moved from outside to inside the organization. For example, although the US Department of Defense has been in existence since 1947, the four services retain their own internal cultures and do not necessarily work effectively with one another.
Governments can also develop procedures to emphasize their priorities and produce enhanced coordination across ministries. For example, the Finnish government developed a program management system that required a new government to select four or five priorities after its election. These priorities cut across existing departments and require the involvement of those departments (Bouckaert et al. 2000;Kekkonen and Raunio 2010). This mechanism does not resolve more general coordination problem but it does permit the prime minister and the remainder of the government to address some high priority policy issues.
Sometimes these procedures can be very simple, such as requiring ministers who want to make a proposal to cabinet to provide their colleagues with 48 h notice, so that the other ministers can assess how the proposal fits with their existing programs. And the budget process should also be seen as a locus for pursuing greater coordination. The budget process may appear to be simply about how much money is going to be spent, but it can be much more. As Aaron Wildavsky once noted, the budget is a statement of government priorities expressed in dollars and cents. Therefore, the annual budget process provides a venue for identifying what rams need to work together more effectively and what may be done to make that happen.
One of the more recent mechanisms developed to pursue coordination is the "czar" (Vaughn and Villalobos 2015). These are officials who are given personal responsibility for a policy domain, and are charged to make that policy work. They generally are responsible directly to the chief executive, and also can utilize his or her political authority to press for better performance and coordination. Although these officials are most visible in the Anglo-American democracies, they have counterparts in a number of other political systems.
In summary, there is no shortage of mechanisms that can be used to promote coordination in the public sector. Any of them can be effective in the right setting, so, therefore, the task for the designer of policy coordination is to understand those settings, and to make the most appropriate selection of instruments. And that designer need not depend upon a single choice. Many governments utilize a wide range of coordination devices, with a need at times to coordinate the coordinators. Although we in the academic community may search for algorithms that provide neat answers to coordination problems, in the foreseeable future, these decisions will depend more on judgment than on policy science.

Is coordination always the answer?
The bulk of this paper has been stressing the importance of coordination for improving the services delivered to citizens. Scholars of public administration and policy have been concerned with failures in coordination since the beginning of these disciplines, and the problems are certainly not solved. But we must also be careful not to assume that coordination is always the answer to the problems that beset government, and, in some instances, too much coordination rather than not enough can be a problem.
The most important reason for questioning the value of coordination is specialization. Governments create ministries of health, or transport, or whatever in order to bring together experts in the field, and to focus on specific policy problems. Specialization is valuable and indeed, some of the reforms associated with the New Public Management have further divided governments and created more specialized agencies that have an even narrower policy focus (Verhoest 2018). Therefore, too much emphasis on coordination may undermine the benefits achieved through specialization.
Second, attempts at coordination may undermine governmental efforts to promote science and the arts. With the possible exception of the Manhattan Project during World War II and other efforts such as Bletchley Park during the war, government attempts to coordinate and manage major scientific efforts tend to be less successful than might be expected given the investment of resources. Science tends to require a great deal of freedom, and attempts to create a more linear and coordinated path to innovation are not as successful as commonly hoped. That said, coordination becomes more valuable in the exploitation of fundamental scientific discoveries.
Third, there may be some rationality in redundancy (Landau 1969). Especially, when governments are engaged in tasks that they want to be sure will be completed, or if they are uncertain about the processes that may be required for success, some redundancy may be beneficial. For example, national security systems may build in redundant intelligence sources and multiple sets of sensors to have the highest possible probability of detecting threats, and also detecting false threats that might cause needless and dangerous actions.
Fourth, too much coordination can be a threat to privacy and to civil liberties for citizens. While we may want coordinated services, having information collection on citizens too well coordinated and integrated can be a threat (Bamberger and Mulligan 2013). The failure to integrate information may facilitate some economic, and even violent, crimes and also may make service delivery for social and health programs more difficult, but that may be the cost that must be paid for the protection of individual rights.
Fifth, competition among organizations, and the resulting lack of coordination, can be used for internal control within the public sector. Accountability is usually conceptualized as hierarchical control of principals over agents, but creating internal competition and using multiple organizations can also promote accountability (Hood et al. 2004). Just as redundant organizations can be used almost as experiments to improve the quality of service delivery so too can they be used as a means of monitoring each other.
Finally, as already noted, hierarchical coordination mechanisms may pose challenges for accountability. If coordinating programs involve blending money from different sources and legal authority from different sources, then central agencies responsible for controlling within the public sector become nervous about their capacity to monitor those organizations and their budgets. While service delivery may be enhanced when coordination structures are developed, some capacity for monitoring may be lost. This can be rectified through the integration of strategic goals into the accountability process 6. Summary: What can we do about coordination? What should we do?
Coordination is a fundamental problem for public administration and policy. It has been recognized as an issue in government for centuries, but continues to vex individuals who attempt to make government work better. Despite numerous attempts to make public organizations work together more effectively, there is still no standardized method for approaching coordination issues, and much of the success or failure of attempts to coordinate appears to depend upon context. Hierarchical methods for coordination may work in some settings but not in others, and that is true for all the options available.
And just as the instruments for addressing coordination problems need to be matched to circumstances, so too does the need to coordinate differ across countries and across policy areas. Some policy domains may work well with minimal attempts to coordinate with others, but others may require substantial policy integration and coordination. Likewise, political systems may emphasize coordination and government more strongly than do others (see Hayward and Wright 2002).
The practical issues for producing coordination are troublesome, but the normative issues involved may be even more difficult. How much effort should be invested in attempting to create coordination, and in what circumstances? Can the resources be better used to deliver the services rather than coordinate them? Although much of the literature on policy coordination treats better coordinated programs this as an unalloyed virtue, in the real world of governing some balancing may be required. The appropriate balance will depend upon a number of factors, but political and professional judgments are required to make the correct decision on coordination.

Disclosure statement
No potential conflict of interest was reported by the authors.