One of the prerogatives that Paul Martin will inherit as leader of the Liberal party and prime minister of Canada is the right to design the formal structure and procedures of Cabinet, including the important role to be played by powerful central agencies like the Prime Minister’s Office, the Privy Council Office and the Treasury Board Secretariat. During his leadership campaign, Martin has been critical of the undue and dangerous concentration of power in the hands of the PM and his key advisors. Apart from his speech on parliamentary reform, he has been vague about possible broader and institutional reforms, but he has clearly left the impression that he favours a less regimented, more open and more pluralistic policy process in which there is less concentration of authority and power in the hands of a small inner circle of political and administrative elites.

Most commentators have welcomed these promises to share power more widely and to create more diffuse sources of policy influence. The prevailing view is that the Canadian political system centralizes power unduly, leading to poor policy choices and a lack of accountability for results. I suggest a countervailing interpretation of prime-ministerial power. I would also recommend that Martin be careful in promising that everyone will get into the act, because the result may be that on the toughest policy issues we will get no action and accountability will become even more blurred.

In formal, constitutional terms, cabinet-parliamentary government is inherently centralized and secretive. Based upon the constitutional principles of  collective and individual ministerial responsibility, power is concentrated in the hands of a few partisan politicians— the prime minister and Cabinet— and they, in turn, are held continuously accountable for the exercise of that power to the public’s elected representatives in Parliament through a variety of mechanisms. This approach to the assignment of political responsibility and accountability differs fundamentally from the presidential-congressional model. Rather than concentrating power and focusing accountability, the American political system disperses power based on the constitutional principles of separation of powers and checks and balances. As a result, the policy process in the United States is usually seen to be more open, permeable, pluralistic, complicated and diffuse in terms of accountability for actions or inactions.

While constitutional principles provide an important starting point, they cannot explain everything that happens at the center of contemporary government. Commentators have long recognized the gap between the theory and the practice of cabinet-parliament government. Over the decades, power is said to have shifted from Parliament to Cabinet and eventually from Cabinet to the prime minister. Most recently, it has become fashionable to talk about prime ministerial power and the dangers to democracy posed by the undue concentration of power in the hands of one individual. Prime ministers have been described as “friendly dictators” and the Cabinet has been portrayed as a mere “focus group” which exists to pre-test prime-ministerial ideas.

Such labels represent good rhetoric but poor analysis. I submit that there are more constraints on the exercise of prime-ministerial power than is popularly assumed. I would also argue in favour of a broader, more pluralistic conception of the policy process than emerges from a debate over whether there is cabinet or prime-ministerial government. A full understanding of the contemporary policy process requires knowledge of the power relationships among numerous actors and institutions, both inside and outside of government. Power depends not only on constitutional principles, institutional structures and the formal procedures used, but also on the wider context and the personalities of the people involved. Clearly, actors depend upon the prime minister, but he also depends on them.

Finally, in an increasingly complicated, turbulent and unpredictable governance environment, formal structures will matter less and political management skills will become even more crucial.

It has usually been assumed that the Cabinet is the forum where the most important decisions in government are taken. Formal constitutional niceties aside, the Cabinet might be said to perform the following list of functions:

• providing information and advice for the PM to perform his responsibilities as leader of the government and leader of the governing party;

• securing agreement among ministers on priorities and arbitrating disputes on issues which extend beyond the boundaries of a single ministerial portfolio;

• providing information and advice to individual ministers on the performance of both their collective responsibilities in cabinet and their individual responsibilities within particular portfolios;

• planning the legislative program and achieving a strategy for securing the passage of bills and expenditures through Parliament;

• providing oversight and coordination of the implementation of policy through the bureaucracy;

• providing a forum for political debate on the political standing of the government, the health of the party and the issues of the day.

Understanding the shifting dynamics of cabinet decision-making is difficult for outsiders because of the practices of cabinet confidentiality and cabinet solidarity. Such practices help the PM and the Cabinet to maintain an outward appearance of unity and they are contributing factors to party loyalty and the maintenance of majority support in the House of Commons.

The charge that prime-ministerial government has replaced cabinet government rests upon an assumption and a number of claims. The assumption is that power resides in one location and is finite in nature. In other words, relationships within the political executive are seen as zero-sum games, in which there can only be winners and losers. The PM has the power and he refuses to share it with others. It is suggested that the prerogatives granted to the prime minister as leader of the governing party and as leader of government mean that he is able to decide policy on all issues in which he takes an interest and, by deciding key issues, he sets the contexts and direction for all other issues facing government. This interpretation makes cabinet a residual organization. Ministers become agents of the prime minister’s will and the Cabinet ceases to function as the key forum where the collective deliberation on policy takes place.

In support of the view that the prime minister has become all-important, commentators point to several sources of his power:

• the prime minister makes thousands of appointments which he uses to promote and to maintain loyalty to his personal leadership;

• the threat of calling an election is a way to block challenges to his leadership;

• control over the agenda of Cabinet and the right to sum up the consensus within cabinet;

• the use of Cabinet committees, informal “kitchen cabinets” and outside advisers, including pollsters;

• the availability of a powerful group of central agencies— the PMO, PCO, TBS and the Department of Finance— to provide countervailing advice to that coming from individual ministers and their departments;

• participation in meetings of first ministers where federal-provincial bargains are struck;

• the media’s focus on the prime minister to the virtual exclusion of everyone else;

• a loyal and disciplined backbench in the House of Commons and a docile Senate that make management of the legislative process predictable compared to other political systems.

I accept that all of these factors make the prime minister the single most important figure within government. However, the impression left of one-person rule is a gross exaggeration. There are a number of constraints on the exercise of prime ministerial power that need to be recognized. In combination these constraints mean that the PM must be careful about how he spends his “political capital,” that is, his reputation for being on the winning side when problems arise. The political capital of the prime minister depends on doing what “significant others” expect, including Cabinet colleagues, the party caucus, powerful interest groups, other governments, the media and the electorate.

The PM works in a number of different worlds. There is the immediate, intimate, intense and introverted world of the inner circle of most trusted Cabinet colleagues, key political advisors, representatives of central agencies, full Cabinet, and certain Cabinet committees. There is also the sprawling expanse of the rest of the government world: individual ministers and their departments, the party caucus, Parliament and the vast agglomeration of non-departmental bodies. A second world in which the prime minister lives is the everyday world of national politics which involves the media, powerful interest groups, provincial governments, the non-parliamentary wing of the party and so on. Finally, there is the shrinking, interdependent world of bilateral and international relations with other countries and international institutions.

Of course, these worlds overlap, intersect and collide at times to make political management of the prime-ministerial agenda very challenging. Once issues reach the agenda of the inner circle of government, the prime minister’s control over them is potentially great and has probably increased during the past three decades. However, in the other two worlds of national and international politics, the PM has probably lost some influence because he is less free to set the agenda of his government. In the more complicated, interdependent and unpredictable policy environments outside of governments there are more surprise events which require a prime-ministerial response, often on an instantaneous basis, the events of September 11, 2001 being a notable example. Furthermore, the national government is increasingly tied down by legal and political agreements with provincial governments and other countries which leave prime ministers less policy room in which to maneuver.

Given the need to deal simultaneously with issues in all three worlds, the PM is forced by circumstance to practice “management by exception.” Time may be the most fundamental limit on prime-ministerial power. No matter how well-organized, well-staffed and energetic he is, no prime minister can arrange to be present when all the “important” decision within government are being made. Sharing the burden of running government necessarily means sharing power. Departmental ministers carry on the great bulk of government programs with little or no direction from the PM.

As governments expanded their scope, there were fewer and fewer new policy spaces to be occupied. Together with the fact of financial restraint, this means that most policies announced by governments today are modifications to existing programs, rather than entirely new innovations. When prime ministers enter office they inherit a legacy of accumulated policies which they usually have no desire or time to change. Prime ministers who become involved with individual departments and programs will soon discover that if operational matters are allowed to come to the top of government, policymaking will move downward because there will not be time for it.

The PM clearly has important levers of influence in relation to his Cabinet colleagues. Most importantly, he is free to appoint, promote, demote and dismiss people from Cabinet. Critics suggest that this leads to Cabinets consisting of “loyalists” and “yes persons.” Two things can be said about such criticisms. First, the critics ignore or underestimate the importance and the difficulty of creating and maintaining a ministry which can reflect and accommodate the diversities of the country. Traditional representation criteria such as region, religion, language and occupation which shaped Cabinet selections in the past have been supplemented in recent decades by gender, ethnicity and race. Often the balancing act of Cabinet selection must be performed within the confines of parliamentary caucus which lacks balanced regional strength. The onus for making the government appear responsive and legitimate to all segments of Canadian society rests with the prime minister.

Secondly, while the PM is “free” to appoint whomever he wants, this does not mean he is surrounded by political nobodies. There are always a number of individuals who will have to be included in the Cabinet because of their status in the party and in the country. There is a well established practice of designating regional ministers within the Canadian cabinet system and these individuals are often strong personalities who cannot be easily pushed around, even by the prime minister. It is hard to imagine that a powerful regional minister like Allan J. MacEachen of Nova Scotia during the Trudeau years or Don Mazankowski of Alberta during the Mulroney years would be excluded from key decisions affectingtheirregions.

The fact that the media focuses its attention on the prime minister gives him prominence but it does not make his Cabinet colleagues any less ambitious. Given the political challenge of finding Cabinet unity and the political strength of certain Cabinet ministers, there is limit to how long a prime minister will want or will be able to keep a contentious issue off the agenda of Cabinet. Smart prime ministers recognize the need to mobilize consent and support for actions of the government and they depend on other ministers to contribute to this process.

The power to dissolve Parliament and to call an election is often seen as a weapon in the hands of the PM. It is, however, very much a twoedged sword. The threat of an election may bring rebellious ministers and backbenchers into line, but no leader will want to face the electorate if his party is openly divided. Control over election timing is mainly an advantage in relation to opposition parties.

In his highly praised book, Governing from the Centre: The Concentration of Power in Canada, Donald Savoie argues that “power has shifted within the centre itself…the prime minister, with the support of advisors in his own office and senior public servants in central agencies, has gained a great deal of power while Cabinet has lost influence.” Savoie places great emphasis on the role of “central agencies” in bolstering the control of the prime minister over the machinery of government. The rise of an extensive central agency apparatus began during the first term of Pierre Trudeau between 1968 and 1972. The intent was, and to a large extent still is, to provide the PM, the Cabinet and Cabinet committees with countervailing sources of policy advice to that flowing from line departments and to help manage the larger volume of business being transacted by the political executive today.

While the principal role of central agencies is to provide support to the political executive, they must also be role models of independent, professionalism for the rest of the public service. As corporate leaders within the public service, central agencies must develop a collaborative approach in their interactions with line departments and agencies.

The constellation of central agencies has changed somewhat over time but the core agencies and their basic roles have remained unchanged. The core agencies are: PMO, PCO, Finance and Treasury Board. The number of central agencies peaked during the early 1980s when there were separate ministries of state for economic and social policy. During the first term of the Mulroney government (1984-1988), the Ministry of Environment temporarily assumed something like the status of a central agency when the government embarked on a “Green Plan” that was meant to be system-wide. More recently, it has been suggested that the Department of Justice plays a centralagency role in ensuring that proposed legislation is “Charter proof.”

During the Trudeau years, the government of Canada became blessed (or cursed) with the most fully developed cabinet support system in the world. That system required ministers to work within an elaborate, specialized and hierarchical network of Cabinet committees and central agencies. Ministers spent endless hours lobbying, bargaining and making recommendations to one another, to central agencies and to deputy ministers serving on so-called mirror committees. There was growing resentment of the powerful influence of the central agencies, especially of the enlarged Prime Minister’s Office, which to its critics resembled the personal advisory system available to presidents of the United States. Also, there was confusion, overlap and rivalry among the central agencies over their respective duties and the philosophy that should guide policy-making. Ministers felt greatly restricted in their freedom to run their own departments and to pursue their own political careers. By the end of the Trudeau era, many informed commentators had concluded that his Cabinet system overestimated the willingness of ministers to act in a collective fashion. Critics of the Trudeau system also argued that leaders of the central agencies often exaggerated the extent to which their organizations could match the range and depth of expertise found in the line departments and were too ready to second-guess the policy advice coming through those channels.

One of the PM’s prerogatives is to decide the internal structures of Cabinet (such as the number of Cabinet committees) and the procedures for generating and dealing with policy proposals. Decisions on these matters are not neutral; they have implications for the centralization or decentralization of power. For example, greater use of Cabinet committees has been seen by some ministers as a way to enhance the power of the prime minister by eroding the role of full Cabinet. There is no “perfect” design for Cabinets. Choices about how power is distributed within Cabinets will reflect the context of government at the time (i.e. the types of issues which dominate its agenda) and the personality of the PM (i.e. his personal leadership philosophy and style).

In this regard, it is interesting to compare the Cabinet model used by Trudeau during the 1970s and 1980s and the model adopted by Jean Chrétien when he became prime minister in 1993. First, in terms of leadership style, Chrétien never had the philosophical and conceptual orientation to politics and government that Trudeau brought to office. Trudeau aspired to make government more rational, innovative, coherent and effective and he believed that this could be accomplished through a series of structural and procedural reforms. Most of the time Trudeau preferred a collegial approach to Cabinet decisionmaking. In contrast, Chrétien saw government as the accommodation of interests; his leadership style was transactional not collegial, and he relied greatly on his personal political skills at managing issues and people within the Cabinet process.

The context is also important in comparing the approaches of the two leaders. In terms of achieving unified direction within his government, Chrétien obtained the benefit of a context which demanded centralized control. From 1994 onward the Chrétien government embarked on a major exercise in deficit reduction called Program Review. Most economic departments saw their expenditures, programs and staff cut by at least 30 percent. Some departments were completely transformed. As the dominant department of the day, Finance used a policy paradigm of deficit reduction and enhanced competition to provide the intellectual coherence necessary to achieve unified direction and fiscal discipline within government. Once Program Review had been completed and the deficit had been eliminated, the normal tendencies of ministers to advocate on behalf of their departments and regions came to the fore again, although in a somewhat muted manner because more realistic expectations of government had taken hold.

Coordination (or more popular today, “horizontality”) has been a perennial quest of practictioners within government. Coordination refers to the harmonization of decisions and activities in the pursuit of an identified goal or purpose. There is both a strategic and preventive purpose to coordination. The strategic purpose involves relating particular decisions to overall goals and to one another. Prevention involves both the avoidance and the resolution of conflicts within and among decisions. Successful coordination also entails the avoidance of gaps in the design and implementation of policy.

As the governance environment changes, the policy and management challenges facing governments in terms of successful coordination are increasing enormously. By governance, I mean the setting of directions and the achievement of change within society. This function has always been shared between governments and other institutions, but the trend in recent decades has been for governments to steer more by remote control and to rely less upon direct intervention. Governance is increasingly interorganizational in character, in part reflecting the fact that we live in a more interdependent world.

In more practical terms, there are a number of recent developments that give rise to new coordination challenges:

• issues are increasingly cross-cutting and do not fit neatly into the traditional ministerial boxes;

• even though governments have somewhat reduced their activities, there is still a high probability that the impacts of one program will spill over into other program fields;

• the existence of federal-provincial agreements on policy and program delivery means sharing responsibility, money, risk and accountability;

• international and bilateral rules of trade and other deals introduce another set of policy considerations;

• governments are relying more on semi-autonomous agencies, commissions and foundations;

• even within traditional integrated departments all the talk (and some of the action) is about decentralization, empowerment of frontline staff and responsiveness to the customer and providing seamless integrated service;

• there is more widespread use of contracting out service delivery and partnership with private sector institutions.

In this kaleidoscopic world of swirling relationships the management of horizontal policy and programs represents a much larger challenge than in the past because governments can rely less on formal structure and commandand-control relationships to achieve coordination. As was seen with the Trudeau model, structures and procedures intended to improve coordination may only complicate the Cabinet system and create their own problems.

How do we prescribe for successful coordination based upon a constantly evolving pattern of interactions and relationships among sets of multiple actors and institutions who are often equal, or nearly equal, in status and power?

The Canadian experience demonstrates that there is as much art as science involved with the design of Cabinet systems. There are no ideal models existing separately from the people who use the system and the circumstances they face. Different patterns of interaction and decision-making have operated during different time periods. The constant factor throughout is that the Cabinet and its related structures are above all the PM’s own instruments for achieving his government’s goals, securing agreement on horizontal actions, allowing for representation, “the challenge function” and due process, and demonstrating responsiveness to changing conditions. These principles confirm that Cabinet is first and foremost a political body for the expression and reconciliation of competing values and interests. Not surprisingly given this central purpose, there is little consensus over the details of Cabinet design.

There is no disputing the fact that prime ministers have important levers of influence that make them the single most important figure in government today. The office is not becoming presidentialized because no prime minister would agree to the checks and balances which a president in the United States faces. Even though the PM is central to the governing process, the institutional norms of Cabinet government mean that issues do not flow as automatically or as quickly to his desk as they do to the desk of the US president. The Cabinet has not become an anachronism or a “dignified part” of the constitution. The cabinet may not deliberate collectively about policy as often as it once did (but even that view of the past is historically suspect), but it continues to contribute to the definition of what actions and inactions are politically acceptable and feasible. To maintain his personal political capital and to preserve the reputation of his government, the prime minister needs on most occasions the uncoerced support of his ministers and to a lesser extent his backbench followers.

The constitutional principle of individual ministerial responsibility and the tyranny of the clock oblige the prime minister to share authority and power with other Cabinet members. Most government activity today consists of running and modifying existing programs in their details— not matters that normally attract prime-ministerial attention. Completely new policy innovations are few. Even with respect to new policies pushed by the prime minister, he ultimately depends upon ministers and departments to carry them out. Within the administrative apparatus of government there is more fragmentation and risk of incoherence due to trends and developments in public management. Many policies today are what Richard Rose calls “intermestic,” that is, they straddle the international and domestic worlds. The prime minister rarely has a free hand to decide such policies.

Also, governments have tied themselves down with an elaborate web of linkages and relationships to other orders of government and private organizations. The proliferation of well financed and articulate pressure groups and more numerous think tanks of various ideological persuasions means that there are more outside sources of policy ideas and demands so that policy analysts within government (including those within central agencies) have far less of a monopoly on the provision of advice to the prime minister.

We have seen during the past three decades the rise of an extensive central agency apparatus. But the influence of central agencies in general, and of particular agencies, has been seen to fluctuate over time depending upon the issues before government and the leadership approach of the PM. At times there have been inconsistencies, contradictions and rivalries among the central agencies, which have weakened their influence. Central agencies remain relatively small compared to the rest of government. There are limits to their capacity to intervene everywhere on behalf of the prime minister. Finally, in the instantaneous world of contemporary politics, driven by the headlines and the deadlines of the media, daily life in the central agencies seems to an outsider to be chaotic, frantic, intense and short-term.

Not all the forces within the Canadian Cabinet system are centripetal in character. There are many centrifugal forces at work as well. In a diverse and open country like Canada we need strength at the centre of government. Strengthening the centre is not necessarily inconsistent with Cabinet or with parliamentary government. Whether we have the balance between centralization and decentralization of power right at any point in time is a subjective, valueladen matter for debate. I doubt that anyone could make the case that Canadians obtain poorer performance and less accountability from their national government than do Americans who live in a political system where power is more dispersed and accountability is more diffused.

Up to a point, the Canadian constitution and related political practices let the PM decide how power will be distributed at the centre. This allows for flexibility in a changing and uncertain world. If a prime minister abuses his power or screws up in policy terms, there is no doubt where blame will be assigned. Paul Martin as prime minister in waiting has made clear his intention to strengthen Parliament, and he has implied that many other actors (provincial governments, cities, aboriginals and citizens in general) will gain influence. However, Paul Martin has to be careful about managing expectations that everyone can obtain what they want from the national policy process and he has to be prepared to defend the crucial role of the centre ofgovernment.