6. Legislative Process and the Mirroring Principle MATHEW D. McCUBBINS At the center of all democratic governments are legislatures. In all legislatures, members compete for access to a variety of valuable resources, such as floor time and committee or cabinet positions. The internal distribution of these re- sources fundamentally shapes the legislative process, and by extension, deter- mines which individuals or coalitions can influence legislative outcomes. In this paper, I argue that, within a given legislature, the distribution of legislative influ- ence tends to mirror the external checks and balances in the polity as a whole. In other words, as Lijphart (1984) has argued, just as polities with little separation of purpose (i.e., with limited diversity of interests and factions) tend to have more unitary governmental institutions than do polities with greater separation of purpose (which tend toward institutions that create separation of powers), so too will internal legislative institutions reflect the separations of purpose and power within a polity.1 This law of organization is referred to as the mirroring principle.2 In making my argument, I consider legislatures generally and use examples from a wide range of parliamentary and non-parliamentary bodies. I argue that many elements of legislatures, as well as the ways that we think about them, are common across most (and perhaps all) legislatures. Thus, a goal of this paper is to present a general analytic framework within which many aspects of the world’s diverse legislative bodies can be considered. I proceed as follows: In the next section, I discuss the nature of legislative resources, and briefly review the various arguments about how they are allocated. The section after that deals with control over the legislative agenda. In the third section, I discuss two cases that illustrate the mirroring principle. I end with a brief conclusion. 1 See Cox and McCubbins (2001) on separation of purpose and separation of power. 2 McCubbins, Noll, and Weingast (1987) coined the term “mirroring principle,” arguing that agency structure mirrors the political forces that create, oversee, and fund the agency. The concept is essentially the same as Ferejohn’s (1987) “structuring principle.” 123 C. Menard´ and M. M. Shirley (eds.), Handbook of New Institutional Economics, 123–147. C 2005 Springer. Printed in the Netherlands. 124 Mathew D. McCubbins 1. LEGISLATIVE RESOURCES AND THEIR ALLOCATION Broadly speaking, resources fall into four categories: The first is legislative time, including caucus time in cabinet or committee, time on the floor (i.e., in the legislature as a whole—sometimes called plenary time), or time in confer- ence committees (in the cases of some bicameral legislatures, such as the U.S. Congress). The second is institutional positions, such as party and committee (or ministry), leadership spots and membership. The third is staff and funding (for either parties, committees, or ministries). And the fourth is legislative out- comes, including budget and appropriation decisions, tax decisions, and vetoes in presidential systems (i.e., the status quo). Given the value of these resources, the method by which they are allocated is of obvious importance. In the case of majority-party-dominated parliaments such as the British House of Commons, it is widely accepted that the majority party controls resource allocation. In many other cases, however—and most no- tably in the U.S. Congress or the Japanese Diet—claims of majority dominance are not as widely accepted (this is also true for U.S. state legislatures). This is also the case for countries in which, due to electoral rules or other factors, legislators have strong incentives to act independently of their parties, or in which presidents possess legislative agenda setting powers (e.g. in Brazil). Ac- cordingly, I spend a good deal of this section reviewing alternative explanations for the allocation of resources. Since these alternative explanations, dubbed the distributive theory of politics, are most fully developed in the literature on the U.S. House of Representatives, I focus primarily on the U.S. case in this section, before returning to a more comparative discussion in subsequent sections. In addition to reviewing these explanations, I outline problems with each of them and then turn to more detailed consideration of the majority-party explanation for resource allocation. Scholars of the U.S. House have suggested a variety of bases for the alloca- tion of resources, including universalism,3 logrolls,4 the regular order,5 the need for policy information, and partisanship. I consider each of these possibilities in turn, beginning with allocation on the basis of either universalism, whereby resources are distributed more-or-less evenly among all members, or by dis- tributive logrolls (Shepsle and Weingast 1987; Weingast and Marshall 1988),6 whereby each gatekeeper7 (be it a committee, ministry, faction, or coalition) 3 “Universalism” implies that all members of the legislature are beneficiaries of distributive policies (see Weingast 1979). 4 “Logrolls” involve two or more legislators agreeing to trade their votes on one bill they care little about in exchange for another bill that is personally much more important to them. 5 “Regular order” refers to the regular rules of procedure in the legislative chamber. 6 The term “distributive” is used because, according to this model, the primary purpose of the House’s internal structure is to make it easy for members to distribute such benefits as government projects, spending, and pork to their constituents. 7 A “gatekeeper” is someone who has access to or a veto over a particular policy or policy area. Legislative Process and the Mirroring Principle 125 and its members get the outcomes that they want in their area of jurisdiction, in exchange for acquiescing to the wishes of other gatekeepers in other areas of jurisdiction.8 For the U.S. Congress, the universalistic and distributive models of allocation are cut from the same theoretical cloth, which views policy making as a logroll among gatekeepers.9 In this view, (1) members self-select onto committees or ministries with jurisdictions of particular electoral value to them; (2) commit- tees and ministries are therefore not representative of the legislature as a whole; (3) leadership positions are distributed automatically via seniority; (4) commit- tees and ministries show deference to one another in cabinet or on the floor (i.e., within each gate-keeper’s jurisdiction, other gate-keepers let them have their way); and (5) members thereby realize gains from exchange by controlling policies that they care about most, while deferring on policies that they care about less. Thus, everyone gets policy outcomes that they want (the universal- ism claim), and members are able to distribute valuable concentrated benefits to their districts. There is some evidence that outcomes are at times universalistic (Weingast 1979; Cox and Tutt 1984), and that floor voting in the U.S. House and Senate became more universalistic during the 1960’s and 1970’s (Collie 1988a). There is also evidence, in U.S. state legislatures, that resources are shared universally: all members receive a committee assignment, travel privileges, resources for mailing their constituents, office space, and staff (in the capitol as well as in their districts). Universal committee representation is not unique to the U.S.; it is found in many legislatures, including most of those in Western Europe (Mattson and Strom 1995). In addition, there is some evidence of distributive logrolls in a handful of policy areas, such as river and harbor improvements and military construction (Ferejohn 1974; Murphy 1974; Weingast 1979; Wilson 1986; Shepsle and Weingast 1987; Collie 1988b; Evans 1994). For example, telecommunications policy in the 1980’s in the Japanese Diet was determined by logrolls between the Liberal Democratic Party’s (LDP) Policy Affairs Research Council (PARC) and committees on telecommunications and agriculture (Noll and Rosenbluth 1995, McCubbins and Thies 1997). In the Costa Rican assembly, small local development projects are allocated by similar means (Carey 1996). While important, however, such logrolls are the exception rather than the rule (Stein 8 In the US House, each committee has authority, or “jurisdiction,” over bills dealing with particular policy issues. This arrangement is similar to the division of policy into various cabinet portfolios in parliamentary systems (Laver and Shepsle 1990; Thies 2001) and is similar to the division of authority among factions in the Japanese LDP (Rosenbluth 1989; Cox and Rosenbluth 1993) and among committee gate-keepers in the LDP’s Policy Affairs Research Council or PARC (Ramseyer and Rosenbluth 1993). 9 Theoretically, this literature has deep roots in transaction cost economics, which focuses heavily on the role of institutions and organization as means of reducing transaction costs involved in both economic and other interactions that occur repeatedly (Coase 1937; Alchian and Demsetz 1972; Williamson 1975; North 1981, 1990; Barzel 1989; Libecap 1989; Miller 1992; Alston, Eggertsson, and North 1996). 126 Mathew D. McCubbins and Bickers 1994), and are not the basis for the allocation of most resources (Browning 1986).10 There are other reasons, both theoretical and empirical, to doubt the distribu- tive model. First, it does not explain changes to the committee system in the U.S. House, or changes in committee autonomy over time—both of which occur in the House across U.S. history.11 It also does not explain a similar evolution of cabinet government in Great Britain (Cox 1987). In addition, members do not freely self-select onto committees, most committees are not composed of preference outliers, seniority is sometimes violated, and committees’ ability to act contrary to the wishes of the floor waxes and wanes over time (Gilligan and Krehbiel 1990, 1994; Kiewiet and McCubbins 1991; Krehbiel 1991; Rohde 1991; Cox and McCubbins 1993). I return to these issues when I discuss partisan allocation below.