Subconstitutionalism.

• Author: Ginsburg, Tom
• Position: Symposium: State Constitutions

INTRODUCTION I. CONSTITUTIONAL THEORY AND THE CONTROL OF AGENCY COSTS A. Theory B. Implications 1. Government structure 2. Rights 3. Amendment C. Conclusion II. SUBCONSTITUTIONAL EFFECTS A. Theory B. Government Structure C. Rights D. Amendment E. Summary III. AMERICAN STATES A. Government Structure B. Rights C. Amendment IV. THE EUROPEAN UNION A. Government Structure B. Rights C. Amendment V. IMPLICATIONS AND EXTENSIONS A. Convergence and Learning B. Moving from Unitary Constitutionalism to Subconstitutionalism CONCLUSION INTRODUCTION

Many nation states have a two-tiered constitutional structure that establishes a superior state and a group of subordinate states that exercise overlapping control of a single population. The superior state (or what we will sometimes call the "superstate") has a constitution (a "superconstitution") and the subordinate states ("substates") have their own constitutions ("subconstitutions"). One can call this constitutional arrangement "sub-national constitutionalism," or, for short, "subconstitutionalism."

Americans understand subconstitutionalism as federalism. The national government controls the superstate; each of the fifty states is a substate. Constitutions exist at both levels. Other states, including Germany, Australia, Austria, Argentina, Brazil, Ethiopia, Switzerland, Mexico, Russia, Venezuela, Malaysia, and Canada, also have federalist or quasi-federalist systems with two-tiered constitutional structures. (1) The integration of Europe has produced a quasi-federalist system. (2) EU members have retained their constitutions even as they increasingly submit to a European government with its own constitution.

When scholars discuss federalism and related forms of decentralization, they typically focus on the constitution of the superstate--the source of the federal structure--and ignore the constitutional aspects of the substates' organization. The justification for federalism is (in modern terms) that some public goods are better supplied at a local level than at a national level because the economies of scale for those goods are not that large, and people can better monitor their government at the local level. (3) This justification is orthogonal to the question of the design of the substate's constitution. To be sure, it implies that the substates must be quasi-independent on some policy dimensions; if they are not, and the superstate ultimately determines local policy, then the system is not federalist. But beyond this minimal level of constitutionalism, many design choices can be made. A subconstitution could contain many rights, or few; it could have a strong system of separation of powers or none at all; it could itself be federalist or not; and it could be easy to amend or hard to amend.

Our interest is the relationship between the superconstitution and the design of the subconstitution. A number of hypotheses are possible. At one extreme, there might be nothing special about subconstitutionalism: the constitutions of substates might reflect the same policy judgments that determine the design of the constitutions of ordinary states. At the other extreme, subconstitutions could have distinctive features. For example, perhaps subconstitutions always mirror the superconstitution. No state in the United States has a parliament. All have three branches of government, modeled after the U.S. Constitution. But there is also a great deal of variation: in the types and number of rights; the procedures for amendment; and the independence of the judiciary, for example.

To our knowledge, none of the work in the voluminous literature on constitutional design directly addresses this topic. Our contribution is to draw attention to the topic and provide a theoretical framework to address it. We use a simple theory that makes a single assumption that distinguishes subconstitutions (that is, the constitutions of substates) from ordinary constitutions: that the superior state in the two-tiered system reduces agency costs that would otherwise exist in the subordinate state. Agency costs refer to the costs that arise as a result of the fact that an agent (here, the government) typically has better information about its actions and their effects on outcomes than the principal (here, the public) does, and can therefore take actions that benefit the agent at the expense of the principal without fear that the principal will learn of that action and punish the agent. When agency costs decline, outcomes improve, and so costly institutions designed to reduce agency costs may be discarded. (4) If agency costs decline when a state becomes a substate, a subconstitution can be weaker than an ordinary constitution is.

Consider a simple example. The U.S. Constitution guarantees a republican form of government for the states. Suppose, then, that the populations of the states can expect the national government to intervene if their republican institutions fail. If this is so, it is less urgent to establish subconstitutions that have strong rules that limit government. By contrast, no foreign states will intervene if the U.S. government loses its republican character, so the U.S. Constitution will need to impose stronger limits on the national government.

If this example can be generalized, it suggests that substates will have weaker limits on government than superstates do. Substates should have weaker government structures (such as separation of powers and federalism), weaker rights, or lowers hurdles to amendment. In the balance of this paper, we lay out the theoretical case and discuss some evidence. Part I describes the economic theory of constitutionalism on which we rely. Part II applies this theory to subconstitutionalism. Parts III and IV discuss evidence from the American states and the European Union. For the sake of brevity, we will not discuss subconstitutionalism in foreign countries such as Germany and Canada in any detail, however, we will refer to some general patterns in those countries. Part V considers implications and extensions.

We offer our theory as a first effort to bring order to a complex and neglected area of constitutional law. We make a number of assumptions that may turn out to be excessively strong, and we acknowledge that, at this point, the evidence is spotty and susceptible to alternative interpretations.

1. CONSTITUTIONAL THEORY AND THE CONTROL OF AGENCY COSTS

   To understand subconstitutionalism, we must first understand constitutionalism more broadly. This Part reviews the literature on constitutions, focusing on features relevant to our account of subconstitutionalism. We follow the rational choice approach of considering institutions in terms of their functions, rather than their values. We define a constitution as a set of rules, superior to ordinary law, that formally binds actors in a political system. Constitutions are typically, though not always, formally entrenched in the sense of being difficult to change. They usually prescribe the process of making ordinary law and define the institutions of government. And they sometimes contain a set of limitations about what that government cannot do, in the form of lists of rights. While there are exceptions, these core features of constitutions are now found in virtually every national constitution in the world.

   As the above description demonstrates, ideas of entrenchment are central to the notion of constitutions. Constitutions are higher law. Their production is associated with founding moments or critical junctures of the state's history. At such points, the ordinary politics of self-interest are sometimes believed to give way to a higher motivation in which fundamental principles are considered and debated. Constitutions are also ascribed a role in forming the polity and creating a shared identity out of disparate parts, thereby contributing to the foundations of the state.

   Why have a constitution? From a rationalist perspective, constitutions are political bargains among important groups in society. The constitution distributes benefits among relevant actors, and also serves to empower and control the agents that produce those benefits. It is this last feature, agency control, that is at the heart of constitutionalism and is the main subject of our analysis.

   1. Theory

      We can begin by imagining a pre-constitutional universe in which each individual participates directly in decision-making about public goods. (5) This would involve extensive discussion and consideration of alternatives before the group made a policy choice on any given matter. Such a system, however morally attractive, faces severe problems of transaction costs and accordingly could operate only on a very limited scale. Constitutions facilitate the hiring of representatives--a government to make decisions about public goods on behalf of the people or other principals. This creates a problem of agency, in which the people must ensure that government acts in accordance with its instructions.

      The relationship between principal and agent is a well-known concept in social science literatures on institutional design. (6) Agency costs may arise whenever a principal hires an agent to perform a given specialized task. Because the principal does not have the same level of information as the agent, there is a risk that the agent might not perform actions in accordance with the interest of the principal. This might be because the agent is acting on behalf of her own interest, or else is captured by (that is, acting on behalf of) a third party. A central task of institutional design is to ameliorate agency costs by aligning the incentives of the agent with those of the principal. Mechanisms for reducing agency costs include devices to screen agents before hiring, to monitor their performance, and to discipline those who do not follow the principal's instructions.

      Even before it was formulated in terms of modern economics, the problem of agency...