STRUCTURE
2. Design of democratic institutions
The main forms of democratic institutions that concern us here are: (i) exec- utive–legislative relations that result from presidential or parliamentary
form of government; (ii) electoral rules, especially the distinction between proportional representation and plurality rule, and the underlying district magnitudes;10and (iii) federalism and bicameral structure.
Presidentialism and parliamentarism
The literature on the consequences of choosing a presidential over a par- liamentary form of government mostly developed during the ‘third wave’
of democratization when the new democracies chose their constitutional structures. In 1978 and later in 1990, Juan Linz stressed ‘the perils of pres- identialism’ and the superiority of a parliamentary government (Linz et al. 1978; Lifz 1990). Although this argument has been challenged and later modified (Horowitz 1990; Shugart and Carey 1992; Cheibub and Limongi 2002), the accepted stylized facts do suggest that parliamentary democracies outperform presidential ones on a number of key indicators, most notably regime survival. Matthew Shugart and John Carey (1992) characterize the distinguishing features of pure presidentialism as follows.
First in presidential systems, the origin and survival of the executive and legislative branches is separate; second, the executive has constitutional authority to execute the laws; and third, the chief executive exercises full control over the cabinet.
There is some disagreement in the academic community on how and whether these core institutions generate divergent consequences under each regime. Furthermore, some scholars point to other differences between regimes that might generate different outcomes. For example, Adam Przeworski and his co-authors (2000) argue that the ‘smaller’ institutions are responsible for the differences in performance between the regimes. To counter this argument, Kent Eaton and David Samuels (2002) introduce two dimensions along which they explore the differences between presi- dentialism and parliamentarism: ‘executives’ unilateral powers’ (such as veto, budget, decree, agenda and other formal powers) and ‘separation of purpose’ between the executive and the legislature (degree to which the executive and the majority of the legislature respond to the same con- stituencies and pressures). Eaton and Samuels go on to show that:
1. presidentialism promotes institutions that strengthen unilateral execu- tive power and separation of purpose between branches;
2. ‘the core institutional differences between regime types are necessary and sufficient causes of differences in political and economic output’
(p. 4);
3. comparable arrangements of non-core institutions (for example, dis- trict magnitude and thus party fragmentation, bicameralism, federal- ism) generate greater differences under presidentialism.
There is a new and growing literature exploring the effect of these vari- ables on policy outcomes in terms of ‘veto players’. The pioneering work of George Tsebelis (1995, 2002) argued that the number of veto players in a system is a function of particular institutional characteristics (separately elected president, federalism, bicameralism) as well as the number of parties. Tsebelis focused on policy change as a dependent variable and argued that as the number of veto players increases, so should policy sta- bility. Josephine Andrews and Gabriella Montinola (2004) have attempted to apply the veto player argument to corruption, contending that as the number of veto players increases, corruption in the system should decrease.11However, Shyh-Fang Ueng (1999) presents a formal model in which, given the same culture, a political institution with veto players to counterbalance its multi-party legislature enables its legislative coalitions to extract larger amounts of bribes from interest groups than one without such counterbalancing veto players.
Issues concerning the salience of these models are not purely empirical.
Recall Eaton and Samuels’s (2002:5) third hypothesis, namely that the same configurations of non-core institutions (for example, district magnitude and thus party fragmentation; bicameralism; federalism) have ‘a greater impact under presidentialism, thus generating additional differences in political output’. Thus, differences in regime performance due to variables such as electoral rules, bicameralism or federalism may be a function of differences in regime type to begin with, since the presence and importance of these variables is associated with presidentialism. Thus, it might be a the- oretical misspecification to lump all the institutions together under the rubric of veto points.
Electoral rules and district magnitude
Consider the following stylized categories of electoral rules: plurality systems with single-member districts and two kinds of proportional repre- sentation (PR) systems – closed and open list (CLPR and OLPR, respec- tively). Under a closed-list system, party leaders rank candidates, and voters only cast votes for parties. Under an open list, voters both select a party and rank candidates given the party’s selection of candidates. In con- trast to PR systems, voters under plurality rule both cast their ballots for specific candidates and elect a single representative from their district of residence. Note that one of the main differences between plurality and PR systems is district magnitude, which refers to the number of candidates elected in a district. Plurality rule usually goes along with single-member districts, which means that a country is divided into a multitude of districts, each of which elects a single candidate. In contrast, PR systems tend to have large, often national, districts, electing the entire legislature in one or
a few districts. Of course, there are differences among PR systems in terms of district magnitude (see Golden and Chang 2006; Persson et al. 2003), but the basic comparison is that under PR district magnitude is larger than under plurality rule.
As Ronald Rogowski (1998) notes, students of politics have disputed the relative advantages of proportional versus majoritarian electoral systems since at least the 1860s (Bagehot 1867; Hare 1889; Hermens 1941;
Rae 1971; Powell 1982; Lijphart 1984). The early stage of the debate focused on why one system should be normatively preferred over another.
The proponents of PR deemed it more encompassing and representa- tional (and therefore inherently more democratic), while its critics argued that it makes any agreement on policy issues difficult to achieve precisely because it includes many actors with different preferences in the policy making process.
Since the 1950s, the debate over the relative merits of PR versus plural- ity rule has focused on the effect of electoral rules on well-defined politi- cal and economic variables, establishing the following stylized facts. PR systems are associated with larger numbers of political parties (Duverger 1954; Rae 1971, Powell 1982; Riker 1982; Lijphart 1984, 1999; Taagepera and Shugart 1989; Cox 1997); party-centered, as opposed to personalistic, political systems (Carey and Shugart 1995; Wallack et al. 2003); more cabinet instability (Powell 1982; Lijphart 1984); higher voter turnout (Beyme 1985); less violence (Powell 1982); greater openness to trade (Rogowski 1989); bigger government deficits and/or higher rates of inflation (Roubini and Sachs 1989; Grilli et al. 1991; later challenged by Lijphart 1984 and 1999, and by Crepaz 1996a, 1996b); corporatism and central bank independence in OECD countries (Anderson 2001); higher price levels in OECD countries (Rogowski and Kayser 2002); and more profit padding by commercial banks (Rosenbluth and Schaap 2003). This generation of literature suggested two apparent trade-offs to be prevalent in PR systems: on the politics side, more representation in exchange for more political instability, and on the economics side, more gains from trade and lower price levels in exchange for higher budget deficits and inflation.
The newer literature moves away from discussing possible trade-offs and notes that PR has worked remarkably well in some types of societies while failing in others. More specifically, Rogowski (1998) notes that PR seems to have performed well in small, open and advanced societies (most notably Scandinavian countries), while performing very badly in larger, less- developed or more competitive systems (Powell 1982; Katzenstein 1985;
Eichengreen 1992). Rogowski then provides one possible underlying mech- anism that differentiates societies in which PR works better from those for
which plurality is more suitable. According to his argument, PR induces the most vote buying and distortionary pork and patronage where voters’ pref- erences are highly volatile, while it works in exactly the opposite direction in societies with stable voters’ preferences. However, more thorough empirical work needs to be done to substantiate these interesting and provocative the- oretical claims.
Federalism and bicameralism
The term ‘federalism’ has been used in different contexts to refer to different institutional arrangements. The most basic definition, however, is that provided by Riker (1964): a state is considered federal if it has at least two levels of government and each level is formally guaranteed (for example, in a constitution) at least one area of action in which it is autonomous. There is some disagreement among scholars as to whether formally federal constitutions are necessary and/or sufficient conditions to classify a polity as federal – Elazar (1987) argues in favor of looking at its constitution to determine whether a state is federal, while Wheare (1963) contends that this is not sufficient and that the practice of government matters just as much. In either case, there are multiple dimensions of federalism that might have policy effects. Daniel Treisman (2002) identifies the following four features of federalism that might affect the quality of government broadly defined:
1. number of tiers of government;
2. decision making and fiscal decentralization – subnational autonomy and taxing/expenditure;
3. electoral decentralization – subnational tiers with locally elected exec- utives; and
4. checks and balances – for example, a regionally elected upper chamber that can block legislation in the lower chamber.
The last point directly interacts federalism with bicameralism. However, bicameralism might have independent effects as well. Some have argued that bicameralism outperforms unicameralism in terms of producing more responsive, accountable and effective governance (Carey 1978; Money and Tsebelis 1992; Tsebelis and Money 1997), while others have noted that bicameral systems are also prone to legislative instability and deadlock due to the excessive checks and balances (Riker 1992; Tsebelis 1995). Also note that some bicameral states use different electoral rules to elect upper and lower chambers, which together with federalism and execu- tive–legislative organization may produce further institutional interaction effects on economic and political outcomes. Let us now examine the
existing theoretical perspectives on the possible links between these insti- tutions and the varieties of corruption described in the previous section.
3. Institutions as incentives for and constraints on corruption: theoretical