Measured by the size of their impacts on the everyday life of most citizens, the legislature, executive, and judicial branches garner a disproportionately large share of scholarly attention relative to that received by the “administrative branch,” the corpus of government composed of hundreds of agencies, commissions, quasi-governmental authorities, and public corporations. These organs of the state are the center of both interpretation and enforcement of the federal government's policies. Leaving aside issues of whether they must necessarily be such, the practical reality is that all interaction between the government and its citizenry involves administrators, whereas a relatively small proportion of government actions are the direct consequence of any decision made by elected representatives or federal judges. While much of public administration is focused on seemingly mundane tasks, the administrative branch is unique in the frequency with which it interacts directly with the citizens. Congruence between government actions and popular will is properly measured relative to administrators' decisions and the explanations offered for these decisions. Viewed from another angle, the congruence between policy outcomes as experienced by the citizens and the policy decisions of the three traditional branches of government depends on the actions and decisions of the administrative branch.

In the first part of this book, we argued that social choice–theoretic results such as Arrow's theorem and the Gibbard-Satterthwaite theorem motivate the study of democratic politics by illuminating the difficulties inherent in accommodating multiple goals within a single decision framework. The theorems tell us that democratic procedures really matter; they must do more than simply select a best outcome, because such an outcome will often not exist. In such cases, it follows, collective choices must be justified, or legitimated, on a basis other than being the unambiguous product of popular will. According to the theory we present and develop in this part of the book, the legitimacy of a collective decision is based on the provision of an explanation (reason, justification, rationale, etc.) for it.

The social choice arguments presented in Part I establish that cycles and intransitivity are endemic to decisions based on the aggregation of multiple criteria (regardless of whether these criteria are directly based on individual preferences). Specifically, the notion of “the best choice” is frequently ill-defined in such settings so that, even after criteria have been duly aggregated, the question of what to choose may not have yet been answered. Such cases are exactly those in which a legitimate choice requires an additional explanation above and beyond the details of the aggregation.

In 1986 the Supreme Court addressed the question of whether political gerrymanders – redistricting plans intended to electorally disadvantage one particular political party – were capable of being, and should be, resolved by the courts. The case in question, Davis v. Bandemer, centered on an allegation by Indiana Democrats that Indiana's 1981 reapportionment plan was drawn by Republicans in order to dilute the Democratic vote and thus violated their right to equal protection under the Fourteenth Amendment. The Court ruled that while an equal protection violation could not be conclusively shown by the appellants in the particular redistricting plan under consideration, such political gerrymandering cases were properly justiciable (or capable of being resolved by the courts) under the Equal Protection Clause. A majority of the Court could not, however, agree on any specific standard for assessing future political gerrymandering claims.

In 2004 the Court revisited the question of political gerrymanders in the case Vieth v. Jubelirer in which Pennsylvania Democrats argued that a political gerrymander had occurred in redrawing Pennsylvania's legislative districts after the 2000 census. However, in this case, a plurality of the Court reversed itself on the Bandemer decision and argued that claims of political gerrymanders were not capable of being decided by a court. Interestingly, the court unanimously agreed that severe political gerrymanders are incompatible with democratic principles.

In this chapter, we describe and explicate the theory of legitimacy in informal terms. However, while our discussion is much less formal than the presentation in the following chapter, a bit of notation will make the discussion here more precise and succinct. Specifically, the basic primitives of our theory are a set of feasible policies, denoted by X, and a set of principles, denoted by P. In addition, when we are discussing a particular principle in P, we will denote it by p. As described in more detail in this chapter, a principle p ∈ P is simply a generalized description of how the various possible policy choices are compared or ranked. The principle is akin to the collective ordering denoted by ≻ in the first part of the book, which we used to denote the output of an aggregative process, f. Thus, a principle represents a method of comparing alternatives that may take many factors into account and that may have been generated via an aggregative procedure. However, in the remainder of the book, we deliberately set aside the question of how the aggregative process of generating a principle worked, the axioms it satisfied, etc. What we take as given is that the principle, as it stands, is the standard by which alternatives are measured and represents the accepted way of choosing alternatives.

Our thesis in this book is simple: choice often requires making trade-offs between various goals. In other words, individuals and groups have various ends and purposes they seek to achieve. For example, from its inception, the founding of the modern federal government of the United States was predicated on the need and desire to aggregate multiple goals:

Of course, it is rare that (say) establishing justice and securing the blessings of liberty will be simultaneously and maximally furthered by the same course of action. Unless such a fortuity occurs, governance – choosing and implementing public policies – requires making trade-offs between these goals. Regardless of whether these trade-offs are explicit or implicit, we have argued that aggregating goals is a ubiquitous characteristic of governance.

Aggregation is the heart of social choice theory. In Part I of the book, we have presented and discussed several seminal results from social choice theory. These results have been misinterpreted by scholars in various ways. The most important misconception about these results, in our minds, is an implicit presumption that the central results of social choice – Arrow’s Theorem and the Gibbard-Satterthwaite Theorem – apply only to methods of aggregating individual preferences.

A key implication of our notion of legitimacy is that legitimacy can require omitting one or more alternatives from the decision sequence. In this chapter, we delve a little deeper into this somewhat perplexing implication. Specifically, we describe its relationship to specific procedural characteristics that can be broadly described as restrictions on scope – that is, standing rules precluding the consideration (or, perhaps, comparison) of one or more alternatives. Such restrictions are common to political and social institutions. In this chapter, we focus first on the U.S. House of Representatives and discuss a key way in which the scope of legislative business is structured in that institution: germaneness. We then turn to the related phenomena of single-subject provisions, a class of rule that limits the scope of statutes.

Before moving to detailed discussions of institutionalized restrictions on the scope of discussion, it is important to note that this chapter focuses exclusively on legislatures. This is by design: the other two chapters in which we discuss real-world institutions focus on the decisions of unelected government officials (judges and bureaucrats). As discussed in Chapter 4, the decisions of popularly elected officials are imbued with a higher level of innate legitimacy than those of unelected officials. Precisely because of this, institutional rules such those discussed in this chapter – rules that both constrain legislative autonomy and are consonant with our theory of legitimacy – are particularly interesting. Within the context of our theory, the germaneness requirement in the U.S.

When considering legislative policymaking, legitimacy is important in two related ways. First, most statutes are generally applicable: there is no guarantee of fit between the requirements of the statute and the facts of each and every covered case. Furthermore, many statutes derive their ability to impact policy outcomes through potentially coercive methods: failure to behave in accordance with the statutory provisions can result in various forms of punishment. In such cases, and as discussed in Section 4.1, the legitimacy of the statute is at best imprecisely gauged through observing compliance. The state's authority, which may be based simply on sufficient strength, can substitute for legitimacy of its policies in eliciting “deferential” behavior by the governed.

Second, legislatures are collective bodies – the statutes they enact are not necessarily transparent in either their consequences or motivations. As the extent of public law and the impact of statutes on everyday life have each grown over the past two centuries, the importance of judicial review – for both verification and explication of the purposes and justifications of legislative action – has concomitantly expanded as well. Thus, while the legitimacy of enigmatic legislative pronouncements is inherently suspect, modern judicial review is theoretically well suited to render both origins and implications of statutes scrutable to the governed.

In this chapter, we formally define our theory of legitimacy and characterize both the sets of policies that may be chosen legitimately and the decision sequences that may be used to legitimate a policy choice. As discussed in the previous chapter, our theory of legitimacy views a legitimate choice as being, for the most part, inseparable from the decision sequence that legitimates the choice. It is arguably not enough to know simply that a policy choice can be legitimated; an observer might reasonably demand to know how the choice in question was legitimately rendered. In this way, our characterization of the decision sequences that legitimate their associated final choice is analogous to a characterization of legitimate procedures.

PRIMITIVES AND NOTATION. Let χ denote the set of all finite nonempty sets of alternatives and, for any set X ∈ χ, let PX denote the set of all asymmetric binary relations (principles) on X. An asymmetric binary relation on a set X, p ∈ P, is a subset of X × X. If (x, y) ∈ p, we write x p y and say that “x is related to y through p.”

The previous chapter details a long-standing debate concerning the field of social choice theory. On one side stands Riker's Liberalism Against Populism, in which the impossibility results are interpreted as proof of the arbitrariness and meaninglessness of populist methods of democratic choice, methods intended to embody a “will of the people” in the actions taken by elected representatives on their behalf. On the other side is a large literature intending to counter these claims by arguing that the impossibility theorems themselves are of little relevance, and perhaps even counterproductive, to the study of democracy. The problem with the theorems, these authors argue, is first that the axioms are unfounded as desiderata of good collective choice procedures, and second, that even if the axioms were well founded, the “impossibility” Arrow predicts is empirically so unlikely to be observed that the theorems themselves are irrelevant.

In this chapter we take up the debate in a more detailed, “claim-by-claim” fashion. In so doing, we argue that the conditions underlying the theorems are in fact both normatively desirable and applicable to many real-world settings. We describe how many of the arguments of scholars who attack social choice theory are based on gross misinterpretations of the meaning of the impossibility results.