«Der Open-Access-Publikationsserver der ZBW – Leibniz-Informationszentrum Wirtschaft The Open Access Publication Server of the ZBW – Leibniz ...»
The second problem is the under-complexity of typologies, which are based only on dichotomies such as unicameral/bicameral systems, federalism/unitarism, or two-party/multiparty systems, and plurality/proportional electoral systems. Tsebelis (1995) has stressed that these dichotomies represent more or less important single elements in more complex institutional arrangements, which then constitute certain types. However, taking these individual elements out of their institutional context can produce misleading results. If we follow Tsebelis’ argument, we face the problem of which institutional characteristics to choose – and, in particular, how they are to be configured to form distinct types. A theoretical criterion is needed for this purpose. Tsebelis himself makes further suggestions in this respect, which we will discuss later. This takes us to the third problem, the choice between a deductive and an inductive procedure in selecting and interrelating structural characteristics.
The most prominent example of an inductive procedure is Lijphart’s (1984, 1999) typology of majoritarian and consensus democracies. Lijphart begins the development of this typology with a description of the so-called ‘Westminster’ model of democracy, with reference to the British system. On this basis, the ‘majoritarian democracy’ type is elaborated.
From the opposing characteristics of majoritarian democracy he constructs the contrasting type of ‘consensus democracy.’ Many authors similarly define ‘presidential democracy’ with an eye on the specific example of the United States. The problem with this procedure is that a democratic regime in a given country can exhibit a mixture of differing characteristics. In Lijphart’s terminology, they can contain both majoritarian and consensus elements. The constitution of a country is not very likely to be consistently designed in obedience to only one principle. However, without an external and
theoretical criterion no decision can be made on the extent to which a specific regime represents a mixture of differing institutional elements or a ‘pure’ type of democracy.
In his final typology of majoritarian and consensus democracy, Lijphart crosses two subdimensions, namely the ‘executive-parties’ and ‘federal-unitary’ dimensions. He obtains these subdimensions through factor analysis of a large number of structural characteristics. However, this permits determination only of empirical coincidence. The empirical co-occurrence of certain structural characteristics in itself reveals nothing about the attributes and effects of these structural characteristics and their combination. For example, it cannot be excluded that a quite specific combination of structural characteristics that is empirically relatively rarely has a strong impact on political performances.
In our study these problems impose two fundamental decisions. First, it must be clear what end the typology serves, and, second, a theoretical-deductive procedure is advisable.
In this procedure, the general standards for typology construction developed and justified by the theory of science provide initial orientation. According to Hempel (1952), a typology should satisfy two main requirements. It should be as parsimonious as possible and it should have the greatest possible discriminatory power vis-à-vis reality. Furthermore, a typology should permit a clear and exhaustive classification of all relevant cases, and it should have the greatest possible explanatory power (Lange and Meadwell 1991).
Discriminatory power and explanatory power can ultimately be determined only empirically. If these two requirements are to be met, a meaningful typology must first be constructed that satisfies the two requirements of parsimony and discriminatory power or unambiguity. Parsimony is required in pursuit of the fundamental scientific aim of obtaining generalizable knowledge. Accordingly, a typology cannot be improved by making it more and more complex but by attempting to establish an optimum balance between theoretically justified abstraction from reality and the most appropriate description of reality.
We attempt to identify fundamental dimensions and distinctions of a democratic regime to provide the basis for constructing optimally parsimonious and powerful typologies and for operationalizing these typologies.
3. Dimensions of a Democratic Regime
Politics is concerned with the production and implementation of decisions with binding effect on everybody. The purpose of this function is to regulate conflicts within a collectivity and to attain collective goals. The ability to make and implement collectively binding decisions is termed political power; power can thus be understood as the decisive medium of politics (Parsons 1969; Luhmann 1974). But political power is a variable element, and this variability depends on two main factors. First, on the way in which power or government is exercised, and, second, how holders of power or rulers are chosen (Loewenstein 1957; Bobbio 1987; Sartori 1987).
Table 1: General Types of Governmental System
Combining these two dimensions with each of two forms produces a simple typology of governmental systems (see table 1). The decisive criterion for selecting rulers is whether they are chosen by the people and thus by the addressees of government, or whether groups and persons appropriate ruling positions by means of the instruments of power at their disposal. The decisive criterion for exercising rule is whether powers are concentrated in one institution (or in one person) or whether they are distributed among several institutions.
Four types of governmental system result: on the one hand monocratic autocracy or polycratic autocracy, and, on the other, monocratic democracy or polycratic democracy. If we replace the Greek term kratos by another Greek word for rule, arché, autocracy and democracy become monarchy (Steffani 1979) and polyarchy (Dahl 1971, 1989). The antique democracy of classical Athens, where power was largely concentrated in the popular assembly, was monocratic democracy. Liberal democracy, in contrast, is polycratic democracy – or, in Dahl’s terms, a polyarchy – where the distribution of power among several institutions is constitutive. Since practically all contemporary democracies are liberal democracies, we concentrate our further analysis on specifying this general type of governmental system.
Table 2: Dimensions of Polycratic (Liberal) Democracy
The specification of polycratic (liberal) democracy in table 2 is concerned with the structure in which the exercise of rule and the selection of rulers takes place. Two structural categories are distinguished: formal structure and empirical or informal structure. Formal structure is the constitutionally defined and thus legally binding set of rules laying down how the two base functions are to be implemented from a procedural point of view. Institutions constitute distinguishable rule complexes with strategically significant functions within the overall institutional arrangements (Parsons 1969; Crawford and Ostrom 1995;
Fuchs 1999). The selection of rulers is determined by the electoral system or electoral law, and the exercise of rule by the codification of institutions and their relations. The latter can be termed governmental system (see table 2). The governmental system and the electoral system together form the democratic regime of a country (Sartori 1994a).
Political scientists postulate a second structural category, which is concerned with interaction between the actors of the democratic regime. This structure therefore constitutes actor constellations that are controlled by informal rules and which can be determined only empirically. They can in so far be called either informal or empirical structures (Easton 1990). One example of such an empirical structure is the party system, which can be described in terms of the number of parties and the degree of polarisation between the parties (Sartori 1976). A party system comes into being through elections; the outcome of elections is influenced but by no means determined by the electoral system (see the relevant arrow in table 2). Just as important are a multitude of changeable societal factors. Actor structures are therefore less defined, more variable, and less predictable than institutional structures. The question is to what extent actor constellations can be considered structures at all and in isolation from the specific level of individual and collective action. It is plausible only if it can be assumed that there are durable constellations of actors to which individual actors adjust and which constrain their behaviour in a systematic manner. To call merely temporary and contingent constellations of actors structures that constrain the action of a given individual actor – as occurs in some veto-player studies – is in our opinion inappropriate.
Regardless of the conceptual problems involved in an informal or empirical structure, this category can be largely ignored for our specific purposes. We are interested in the quality of different democratic regimes that are constitutionally defined and can thus be intentionally designed. Given this focus, the formal (constitutional) structure of the democratic regime is the central concern of our analysis. The following section discusses the basic options for constructing typologies of democratic regimes.
4. Typological Approaches to Democratic Regimes
The fundamental criterion for describing and distinguishing democratic regimes is the distribution of power, with the two poles concentration and dispersion of power. In the case of a democratic regime it can only be a question of relative proximity to one or other of the two poles, for a minimum of power distribution among different institutions is a defining characteristic of liberal democracy (see table 1). The distribution of power was the aspect that dominated the constitutional debate in the 18th and 19th centuries – the terms then current were the separation or division of powers – and is still the prime point of reference for almost all typologies of democratic regimes. In a democratic regime power is distributed among different governmental institutions.
In this conceptualisation, the democratic regime is largely equated with the democratic system of government. The electoral system as the institutional mode of selecting rulers is taken systematically into account by none of the typologies and indices we have examined.
In some indices it is merely one (additive) characteristic of the governmental system. Sartori (1994a) considers the electoral system and the governmental system – as the two base functions of liberal democracy: selection of rulers, exercise of rule (see table 2) – to be the two fundamental dimensions of a liberal democracy. But he proposes no typology and no index constructed on this basis.
The two fundamental institutions of the governmental system of liberal democracy are the government and parliament. Each of these institutions has its specific function in the exercise of power. The relationship between the two is the point of departure for one of the oldest and simplest typologies of democratic regimes: the distinction between presidentialism and parliamentarism (Loewenstein 1957; Verney 1959; Steffani 1979; Shugart and Carey 1992; Weaver and Rockman 1993; Sartori 1994a, 1994b). This distinction is initially somewhat misleading, for it suggests that presidential systems are not parliamentary systems. This is, of course, not the case. As in all liberal democracies, parliament has a ‘significant’ function in the presidential system, too (Steffani 1979) – in legislation, in the budget, and in controlling the government. The difference between presidentialism and parliamentarism lies in the constitutionally defined relationship between parliament and government.
Two paired concepts are used to describe this relationship, which mean substantially the same: autonomy (presidentialism) versus dependence (parliamentarism) and separation (presidentialism) versus integration or fusion (parliamentarism). The meaning of these paired concepts is generally defined operationally by stating certain constitutional characteristics. Shugart/Carey (1992) and Sartori (1994a, 1994b) describe a purely presidential system in terms of the following characteristics: 1. the president is both head of state and head of government (monistic executive) and he is also directly elected by the people; 2.