In framing a government which is to be administered by men over men, the great difficulty lies in this: you must first enable the government to control the governed; and in the next place oblige it to control itself. (James Madison, Federalist No. 51)
‘All politics is local’, according to the late Massachusetts Democrat and House Speaker Thomas P. (‘Tip’) O’Neill. Nowhere is that adage truer than in geographically based representative democracies.
Direct democracy, whereby political decisions are taken by the enfranchised polity as a whole, is certainly not unknown. That method of collective decision-making is utilized in many townships in the United States and some municipalities in Switzerland. Voter referenda and citizen initiatives for such purposes as approving bond issues and amending constitutions are common features of US state and local governmental processes (Sass, 2001: 157); referenda are likewise used extensively in the EU for important constitutional issues. Moreover, the institution of direct democracy is likely to be exploited more frequently in future as changes in information technology continue to lower its costs. It is nevertheless true that, because the cost of group decision-making rises sharply with group size (Buchanan and Tullock, 1962; Olson, 1965), in most democratic societies ordinary citizens delegate decision-making authority to politicians they elect to represent them in a legislature or parliament.
The legislative assemblies that, through the enactment of laws, conduct the ordinary political business of representative democracies exhibit astonishing diversity around the globe. Buchanan and Tullock (1962: 223-224) identify four key constitutional variables that shape the form and functioning of these democratic institutions. First are the rules for electing the members of the legislature. Are legislative representatives elected by a plurality, by a simple majority, or by a qualified (say, two-thirds) majority of the enfranchised citizenry? Second are the rules that define the basis of legislative representation. Do legislators represent functional interests (e.g., labor, agriculture, property), do they have geographical constituencies, or is representation in the legislature instead based on legislators’ political party affiliations and proportional to party strength in legislative vote totals? Are legislators elected at large, or is the basis of representation apportioned amongst single-member constituencies? Third are the rules that determine the degree of representation (i.e., the number of voters per representative) and, hence, the size of the legislative assembly. Fourth are the rules that govern action by the legislature itself. Can a simple majority of the members enact laws, or is a more inclusive majority required?
The constitutionally specified number of legislative chambers also belongs on this list. Is the legislature uni-cameral, consisting of a single body, or is it bicameral, comprising two chambers, an upper house (e.g., a Senate, a House of Lords, or a Bundesrat) and a lower house (e.g., a House of Representatives, a House of Commons, or a Bundestag)? In any case, these four (or five) variables interact to determine the constellation of interests that get represented effectively in the legislature. They do so, chiefly, by affecting two types of costs the members of the legislative assembly face. The first of these is decision-making costs, the direct and indirect (opportunity) costs of gathering information, negotiating vote trades, and achieving the level of agreement required for taking legislative action. Other things being the same, decision-making costs tend to rise as the size of the legislature increases, as the interests of the constituencies represented become more heterogeneous, and as legislative voting rules become more inclusive. The second category of costs is external costs, the reduction in utility the members of any random minority can expect to have imposed on them because of actions taken by any random majority. Increases in the degree of representation, greater homogeneity of constituents’ interests, and voting rules that are more inclusive tend to lessen these costs.
To illustrate, consider a unicameral legislature and a bicameral legislature yielding equal degrees of representation (i.e., having the same number of constituents per representative). If legislative action can be taken by simple majority votes in both cases, it follows that, in the absence of a vote cycle, similar legislative outcomes will emerge under unicameral and bicameral structures unless the members of the latter’s two chambers have different bases of representation (i.e., serve different constituencies). With the same total size and the same basis of representation, the cost of obtaining a legislative majority is the same; the fact that one legislature has two houses is purely cosmetic. Now suppose that the basis of representation of the bicameral legislature’s upper house differs from that of its lower house, and that legislative action requires a simple majority vote in both chambers. Given that a majority of the representatives of two constituencies, dissimilar in their respective interests, must vote favorably before the legislature acts, the cost of obtaining a legislative majority is plainly higher than it would be in a unicameral legislature of the same total size, the members of which have the same basis of representation. By increasing the cost of legislative agreement, bicameralism serves as a brake on an activist legislature.
The costs of enacting legislation in a bicameral legislature also rise with greater disparities in chamber sizes. If chamber sizes are unequal and the legislature’s decision-making costs increase at an increasing rate, the cost of buying one more vote in the larger chamber will exceed the savings from buying one less vote in the smaller chamber. Assuming independence between the two chambers, it follows that the total cost of securing a bill’s passage will be greater the more unequal are the sizes of a bicameral legislature’s upper and lower houses (McCormick and Tollison, 1981: 44-45). These observations also suggest that legislative structures trade off in predictable ways with legislative voting rules: other things being the same, outcomes in uni-cameral legislatures will approximate those of bicameral legislatures only if unicameral legislatures operate by voting rules that are more inclusive than simple majority.
Total legislature size, by itself, has an ambiguous effect on the cost of legislating. On the one hand, by reducing the average number of constituents per member of the legislative assembly, larger legislatures tend to lower voters’ costs of monitoring their elected representatives. On the other hand, because each member of a large legislature will have less influence on the legislative process, such monitoring will pay fewer dividends. By diluting the value of each legislator’s vote, large legislatures also lower the price at which interest groups can buy influence. The effect of legislature size on the price of influence is offset to some extent by the logic of collective action (Olson, 1965), which suggests that it will be more difficult in large legislatures for a majority of the members to reach the agreement necessary to enact legislation. Votes are cheaper in large legislatures, but to assemble a majority coalition more of them must be bought.
Citizens delegate decision-making authority to their elected representatives in order to reduce their own decision-making costs. Insofar as these representatives are utility-maximizing human actors, that delegation creates a principal-agent problem for voters: legislators will pursue their own self-interests to the extent allowed by given political constraints. The seeking of personal wealth, the consuming of the perquisites of legislative office, and the angling for post-legislative-career job opportunities supply incentives for legislators to ‘shirk’, either by voting against their constituents’ interests (Bender and Lott, 1996), or by indulging their own ideological preferences (Kau and Rubin, 1979, 1993; Rubin, 2001).
Perhaps the most salient personal goal of an elected representative is to win reelection. Because the mass of voters is rationally ignorant of the impact most of the business of the legislature has on their private wellbeing (Downs, 1957) and, moreover, because voters face serious free-rider problems in monitoring the behavior of their elected representatives and in removing them from office, the members of the legislative assembly have little motivation to advance the interests of their constituents as a whole. Bolstered by the well-known advantages of incumbency, legislators will instead cater to the demands of special-interest groups that are well organized and willing to deliver political support (i.e., campaign contributions and vote blocs) in return for promises of favorable legislative treatment.
‘Pork barrel’ politics is the predictable outcome of legislators’ tendencies to advance the parochial interests of relevant electoral coalitions, ‘however unimportant the interests may be from a national standpoint’ (Posner, 1969: 83). Particularly in a geographically based representative democracy, the members of the legislative assembly will have strong incentives to support programs and policies whose benefits are concentrated on the individuals and groups important to them politically, but whose costs are distributed diffusely over the polity as a whole. While the representatives of the constituencies who will bear the tax burden of financing projects from which they derive little or no benefit would otherwise vote against the proposal, they have pet projects of their own designed to benefit their political supporters at the taxpaying public’s expense. Although no one pork barrel project would pass if voted on in isolation — because its costs summed over all voters exceed the value of its narrowly focused benefits — there are obvious gains to be had from legislative vote trades. These gains can be exploited through bilateral or multilateral agreements in which a majority of the legislators each promise to vote for the others’ projects (‘logrolling’), or by packaging the proposals together in an omnibus spending bill that each member of the majority is obliged to support in order to secure approval for his own project. Explicit or implicit vote trades facilitate the playing of a legislative negative-sum, prisoner’s dilemma game from which no one member can withdraw and hope to retain his seat.
The necessity of monitoring compliance with logrolling bargains helps explain why most legislative voting takes place near the end of the legislative session (Crain et al., 1986). (Ex-post settlements are another means of enforcing legislative vote trades.) The ordinary business of the legislature, which consists primarily of brokering wealth transfers among various identifiable groups within the polity, is facilitated by a legislative committee system that assigns responsibility for overseeing defined policy areas (agriculture, defense, commerce, and so on) to subsets of the legislature’s membership. Legislative committees, who members have immediate political interests in the policies they oversee, help control the legislative agenda (Leibowitz and Tollison, 1980), thereby averting vote cycles, monitor the executive branch agencies that implement legislative policies (Weingast and Moran, 1983), and serve as proving grounds of party loyalty (Crain, 1990). Public choice analysis has been applied fruitfully to these and other institutional features of legislatures (Tollison, 1988).
When the Founding Fathers of the US constitutional republic designed the national legislature, they were well aware that its members would be subject to the influence of politically powerful special-interest groups, or ‘factions’. Although James Madison strongly opposed the Constitutional Convention’s ‘Great Compromise’, which granted the states equal representation in the Senate, he later wrote eloquently in defense of the separation of powers embodied in a bicameral Congress. Composed of a lower house, its members having relatively short terms (two years) and its seats apportioned on the basis of population, paired with an upper house, its members having relatively long terms (six years) and its seats apportioned equally among the states, ‘no law or resolution can now be passed without the concurrence, first, of the people, and then, of a majority of the States’ (Federalist No. 62). Recognizing the value of impediments ‘against improper acts of legislation’, Madison clearly understood that what might be called the industrial organization of the legislature (Weingast and Marshall, 1988) — its structure and operating rules — has an important influence on legislative outcomes.
