Delegation in Contemporary Democracies

Fabrizio Gilardi, Dietmar Braun, Fabrizio Gilardi, Dietmar Braun

• Publisher: Routledge
• Year: 2006
• Print ISBN:
  9780415543569
• eBook ISBN:
  9781134261413
• Topic: Politics & International Relations
• Subtopic: Comparative Politics
• Index: Politics & International Relations

Delegation is an ubiquitous social phenomenon linked to the growing differentiation of modern societies. Delegation is one of several different modes of organisation that exist to make collective action successful, but has been overlooked and under-researched.

Using a rational choice institutional analysis and principal agent models, this book brings literature on delegation to bureaucracy, electorate to legislature to government within representative democracy together with literature on new forms of delegation such as non-majoritarian institutions, to provide a more complete and synthetic analysis of delegation in political systems.

With a broad and comparative approach, this is an important volume for advanced students, researchers and professionals concerned with delegation in the areas of public policy, public administration and democratic theory.

1 Introduction

Dietmar Braun and Fabrizio Gilardi

Delegation is a ubiquitous social phenomenon linked to the growing differentiation of modern societies. With the division of labour and specialization, the multiplication of ‘social circles’ (cf. Simmel) and the genesis of functional systems and subsystems, the power and capacity of individuals to realize objectives on their own account fall into question. The realization of objectives depends to an increasing degree on what other individuals do and, thus, is subject to collective action. Different modes of organization exist to make collective action, which is often plagued by ‘opportunism’ and non-coordination, successful: hierarchy, markets, networks and delegation. While the former modes have been elaborated in detail in the literature, delegation has remained somewhat the stepchild of political theory.

While hierarchy depends on command, markets on unplanned coordination by prices, and networks on the will to cooperate on an equal base, delegation can be considered as an ‘extension of self’ (Coleman, 1990) by transmitting authority in the form of property rights to someone who is considered to have capacities that the ‘delegator’ – or in terms of rational choice theory, the principal – doesn’t have but wants to make use of in order to improve his well-being. Delegation is, therefore, a form of collective action where we find a principal (the ‘delegator’) and an agent (the ‘delegated’) who combine forces in order to realize the objectives of the principal. In doing so, the sphere of action of the principal is extended. In contrast to hierarchy, delegation does not use commands but contracts and incentives, which the agent voluntarily accepts. In contrast to the market, the coordination in delegation is conscious and based on explicit contracts. While actors in networks deal with each other from an equal base of power, the delegation relationship is asymmetric; the agent is supposed to work for the principal to realize their aims.

Forms of delegation have been a subject in political science for quite some time, though a profound theoretical controversy on the main components and rationales of delegation as a mode of collective action is lacking.

The most obvious political phenomenon is democratic representation, which was initially discussed in Montesquieu’s and Burke’s treatises on the relationship between the citizen and his representative in parliament. Burke demanded – in opposition to Rousseau, who wanted the direct binding of the representative to the will of the citizen – the delegation of authority from the citizen to the deputy, who then deliberates with other deputies in order to define the ‘common will’ (Burke, 1863). More recently and more in detail, Hannah Pitkin has elaborated on the different forms of democratic representation and their rationales (Pitkin, 1961). Her work can be considered as one of the early contributions to a then not yet existing theory of delegation as representation.

In the 1920s, Max Weber and his disciple Robert Michels attacked another obvious delegation phenomenon, i.e. the role of party leaders and their followers. The ‘iron law of oligarchy’, pronounced by Michels, demonstrates one of the main problems of delegation, the seemingly inevitable growing autonomy and estrangement of the delegated from the delegators (Michels, 1911).

One can consider in general the political literature on elites and their relationship with the ‘masses’ and citizens respectively, dealt with for example in the work of classical authors like Pareto, Mosca, Schumpeter and C. Wright Mills and modern authors like Putnam (Putnam, 1976) and Bourdieu (Bourdieu, 1981, 1984), as a potential contribution to a theory on delegation.

Whenever – and this is not necessarily only the case in democracies – citizens transfer some of their rights of sovereignty to delegates on the political level, a delegation relationship is established.

While this literature discussed delegation as representation, delegation may also be found on the executive side of policy making. An obvious example would be the role of bureaucrats in public policy making. Though Max Weber was certainly among the first to have scrutinized this phenomenon, his discussion does not analyse the relationship between politician and bureaucrat in terms of delegation. According to his studies, bureaucrats are engaged in a hierarchical relation with politicians and have no leeway for action. Hierarchy is, however, different from delegation as stipulated above. It was Niskanen in the 1960s who made it clear that bureaucrats have their own will and room to manoeuvre (Niskanen, 1971). The relationship should not be considered hierarchical. When politicians transfer some of their property rights to bureaucrats, they must be aware that the logic of bureaucrats’ actions can fundamentally differ from pure execution of governments’ orders.

Corporatism, especially in the form of ‘state corporatism’ (Schmitter, 1979), is another example of delegation on the executive side. The delegation of important functions of the political system such as the organization of industrial relations or social security to corporate actors in society clearly manifests the willingness of policy makers to transfer some of their property rights to societal actors in order to reduce transaction costs and forgo information problems.

The more recent studies on the role of ‘independent agencies’, most notably in regulatory policies, prominently put forward by Majone’s considerations of a shift from the redistributive to the regulatory state (Majone, 1997), can be regarded as a last domain where political scientists are boarding the ‘train of delegation’. Independent agencies can be considered as a particular form of execution, similar to bureaucratic action but with more discretion than is usually attributed to the bureaucrat. This raises particular problems that are dealt with in the literature.

Although we find a large number of political phenomena that can be scrutinized under the denominator of ‘delegation’, a theory of delegation embracing and synthesizing all these different approaches is still lacking. Recently, principal–agent theory, a strand of rational choice theory, has become prominent in studies dealing with delegation and has become a unifying paradigm in the building of a theory of delegation. What are the basic components of principal–agent theory with respect to delegation?

Principal–agent theory

Principal–agent theory is based on an economic theory of choice and behaviour. It assumes that actors are interest-maximizing and opportunistic. Although phenomena of sympathy and identification can occur within an agency relationship (Coleman, 1990: 157–162), rarely is there complete harmony of interests between the principal and his agent. This means that most of the time there is, at least potentially, a conflict of interests between these two actors, which is frustrating because the very raison d’être of the agent is that of fulfilling the principal’s interests. Given this configuration, the principal’s problem is motivating his agent. In particular, the problem is to establish an incentive structure that leads the agent to maximize the principal’s interest. In fact, since actors behave opportunistically, an immediate implication of the conflict of interests is that the agent will systematically try to maximize his own interest instead of that of the principal, in other words, the agent will tend to shirk.

Another important assumption is that information is asymmetrically distributed between the two actors, typically being in favour of the agent, who knows more about himself than the principal does. This causes two main problems for the principal: two different kinds of opportunistic behaviour, one known as adverse selection (or ex ante opportunism, or hidden information) and the other as moral hazard (or ex post opportunism, or hidden action). Adverse selection occurs whenever the principal cannot be sure that he is selecting the agent who has the most appropriate skills or preferences. Moral hazard occurs whenever the agent’s actions cannot be perfectly monitored by the principal. These are general problems inherent not only in agency relationships but more generally in all contractual and hierarchical relationships (Moe, 1984: 755).

These problems could be mitigated if the agreement between the principal and the agent could be fixed in perfect, complete contracts. Such contracts, however, would have very demanding requirements (Milgrom and Roberts, 1992: 127–132). Due to the fact that actors are only rationally bounded, actual contracts will always suffer from incompleteness. In actual contractual relations, ‘contingencies inevitably arise that have not been planned for and, when they do, the parties must find way to adapt’ (Milgrom and Roberts, 1992: 128). This has the effect of worsening the problems of opportunistic behaviour rather than mitigating them.

Given these informational and motivational constraints, the principal’s problem is to prevent the agent from shirking. It is thus not surprising that scholars, especially in economics, have mainly directed their efforts to understanding what kind of incentive structure the principal should (and, since he is assumed to be rational, will) set up in order to minimize the agent’s shirking. Hence, an extensive amount of the economic literature deals with optimal incentive structure issues (e.g. Mirrlees, 1976; Glover, 1994). In political science, however, principal–agent models have also been used to investigate more substantial issues such as delegation in political settings. In particular, on the basis of principal–agent theory both legislative and executive relationships in parliamentary democracies have been subject to a reconceptualization in terms of delegation.

The chain of delegation in parliamentary democracies

Democracies have been conceptualized as chains of delegation by Kaare Strøm (2000, 2003; Strøm et al., 2003). The idea is that the influence of democratic principals, i.e. citizens, on policies is shaped by the different formal institutions that separate them from the final decision. These institutions can be seen as delegation steps, as principals give instructions to agents and thus delegate to them decision-making authority. In the first place, citizens delegate authority to their representatives, who usually sit in parliaments, and in presidential systems also in the presidential chair. Then, in parliamentary systems, the parliament delegates some of its authority to the government, which, as a whole, further delegates to the individual ministers. Next, the government typically delegates policy implementation to the bureaucracy. Strøm thus identifies four main links in the delegation chain. As will be shown below, delegation to independent agencies can be seen as a fifth step, which seems to be neglected in this literature.

The ideal-typical chains of delegation in parliamentary and presidential democracies can be differentiated on the basis of the complexity of the chain. Parliamentary democracies are characterized by the singularity principle, according to which each principal employs a single agent (or a set of non-competing agents), as well as by indirect delegation, as few agents (members of parliament) are directly elected by citizens, and most agents (ministers and bureaucrats) are indirectly appointed, often through several stages. Presidential democracies, on the other hand, feature a more complex chain that, however, ensures a more direct link between citizens and agents. Most notably, the government is directly elected by citizens in presidential systems, but not in parliamentary ones. Different institutional arrangements lead to different consequences, and thus agency losses and control mechanisms follow from the structure of the delegation chain. In particular, the fact that in parliamentary democracies (in contrast to presidential systems) there is less reliance on checks and balances implies that ex post controls are weak and ex ante controls are strong. The more important role of parties in parliamentary democracy leads to effective screening devices that mitigate adverse selection of agents.

Equipped with this overview of the main features of the delegation chain, we can now turn to the single steps.

Delegation from voters to elected representatives

The first step of delegation is from voters to legislators. Mitchell (2000) argues that this agency relationship is framed by two sets of rules, namely party methods of candidate selection and electoral laws, which are best thought of as control mechanisms intended to control agency losses. Electoral laws and candidate selection by parties constitute ex ante means to avoid adverse selection, whereas ex post controls are based mainly on monitoring and reporting by third parties such as interest groups and the press, but also on institutional checks. More specifically, agency losses are more likely to occur when legislators’ careers do not directly depend on the preferences of voters, because in that case incentive compatibility between the former and the latter tends to be lower. Müller (2000), on the other hand, considers legislators as agents not of voters, but of the extra-parliamentary party organization. The question here is how can the party control its representatives standing for or already in public office through the standard ex ante (contract design, screening and selection) and ex post (reporting requirements and monitoring, institutional checks) control mechanisms? (Müller, 2000: 322–329).

Delegation from legislators to the executive

The second link of the chain of delegation in parliamentary systems is that between members of parliaments and governments. Saalfeld (2000) identifies two key elements of legislative organization that are likely to affect the nature of the agency relationship between parliament and government. First, he stresses that the delegation process from members of parliaments to the government is mediated and controlled by political parties. Second, he argues that committee specialization is a response to the growing information asymmetry between parliament and government, but is also a new source of agency problems, namely between committee members and non-members.

Delegation from the governmental cabinet to individual ministries

The third link of the parliamentary chain of delegation involves cabinet and ministers. The agency relationship deriving from this link, however, is fundamentally different from the others. In fact, in this case ‘the principal is made up of its own agents’ (Andeweg, 2000: 377); the government delegates to ministers, yet it is also composed of them. This, it is argued, opens up significant potential for agency losses. Andeweg (2000) makes the point that the nature of the agency relationship between the government and its ministers varies along two dimensions of government decision making, namely whether it is organized hierarchically or collegially, and whether it is strongly departmentalized (that is, each ministers enjoys considerable autonomy in its own legislature) or, rather, decisions are taken collectively. Agency losses are more likely when ministers are more autonomous and when the government is organized collegially. In that case the prime minister does not have the right of supervision over the ministers’ activities, and the government is departmentalized, so that each minister has nearly complete autonomy in his own legislature. Conversely, agency losses are less likely when the government is hierarchical and the decisions are taken collectively.

Delegation from the government to the bureaucracy

Delegation to bureaucracy is the fourth link in the chain of delegation and, especially in the United States, was the first area where principal–agent models were applied in political science. American scholars have spent a good deal of time and effort in debating the extent to which the bureaucracy is autonomous from Congress and/or the President (for an overview, see Moe, 1987; Huber and Shipan, 2000). Consistent with the assumption of the principal–agent model, they have looked for and identified several mechanisms through which politicians can minimize bureaucratic drift, i.e. the difference between the policy chosen by t...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Illustrations
• Contributors
• Series editor’s preface
• Acknowledgements
• 1 Introduction
• Part I The standard chain of delegation
• Part II The next steps in delegation