«Der Open-Access-Publikationsserver der ZBW – Leibniz-Informationszentrum Wirtschaft The Open Access Publication Server of the ZBW – Leibniz ...»
the tenures of president and parliament are fixed and there are practically no mutual powers of removal or dissolution; 3. the president determines the composition of the government and the decision-making activities of the government. Although in principle presidentialism-parliamentarism typologies describe parliamentary systems in complementarity to these characteristics, they are less clearly defined. The most important characteristic of parliamentarism is the choice of the head of government by parliament and the responsibility of the government to parliament. The interdependence of government and parliament is most marked when parliament removes the government by a vote of no confidence, and the government can dissolve parliament (Loewenstein 1957). Both possibilities imply that there can ultimately be no fixed terms for parliament and government. These constitutional characteristics of the parliamentary system can, however, be frustrated if the government is formed by one party over a long period and can rely on a majority in parliament. The institutional design of the governmental system imposed by the constitution or by constitutional equivalents is then subordinated to the constellation of the party system. The political outcome of this subordination is the extreme predominance of the government over parliament, as has been the case in Britain.
The archetype of a presidential system is the United States. The advantages and disadvantages of this form of government are often discussed on the basis of this example. The main reason why the fathers of the American constitution established a ‘system of separation of powers’ (Weaver and Rockman 1993: 2) was to prevent tyranny and any abuse of power. In the current discussion on the effects of presidentialism – and thus of parliamentarism as the contrasting type of regime – this aspect has tended to be relegated to the background. Attention is focused on the efficiency of decision-making processes and the effectiveness of the intended policies (a compilation of this discussion is provided by Weaver and Rockman 1993; see also Schmidt 1995 and Tsebelis 1995). It is repeatedly assumed that, because of the relatively marked autonomy of the government (president) on the one hand and parliament (Congress) on the other, the presidential system has at least three major disadvantages over the parliamentary system: higher transaction costs, less coherence among individual policies, and less capacity for political innovation. This assumption, astonishing in the light of American history, would first of all need to be empirically verified by comparative studies – existing findings tend to be contradictory. But even if these assumptions are taken as premises, they raise two problems. First, the parliamentary system cannot without further ado be described in contrast to the presidential system as a ‘system of fusion of powers.’ As far as the distribution of power criterion is concerned, there is considerable variance within existing parliamentary systems. Second, the mutual independence of executive and legislature underlying the presidentialismparliamentarism typology is only one of the constitutional characteristics of power distribution (others include bicameralism and federalism). That the governmental system of the United States is empirically characterised not only by presidentialism but also by federalism and bicameralism does not necessarily mean that this is true of other presidential systems – and this is observably not the case.
The fact of a plurality of institutional characteristics that play a role in the distribution of power is taken up by the veto-player approach and placed in a quite independent theoretical context. This approach refers explicitly to classical ideas of the separation of powers, reformulating and formalising them with recourse to theorems and methods taken from the rational-choice paradigm. Formalisation is concerned primarily with the relationship between independent variables (veto-player indices) and dependent variables (measurements of ‘policy outcomes’ or ‘policy innovation’). In our context, however, we are interested only in ideas developed in the framework of the veto-player approach on distinguishing types of regime. The best-known variant of this approach is that of Tsebelis. One of his definitions of the veto player states: ‘a veto player is any player – institutional or partisan – who can block the adoption of a policy’ (Tsebelis 1995: 305). An institution counts as a veto player only if it has a formal power of veto. The number of institutional veto players and their power of veto is defined by the constitution. In contrast, Tsebelis (1995: 304) sees the ‘partisan veto player’ as endogenously specified by the party system and government coalitions (this corresponds to the empirical structural categories shown in table 2).
Tsebelis’ dependent variable is ‘policy innovation,’ and both categories of veto player restrict the action of policy-process actors. In this way they determine ‘policy innovation.’ We are interested in institutional veto players. Tsebelis (1995) and Tsebelis and Money (1997) stress the particular relevance of two constitutional characteristics that define different institutional veto players: presidentialism and bicameralism. Federalism is also discussed. However, it is not clear whether these authors see it as contained in bicameralism or as a separate constitutional characteristic. The question Tsebelis explicitly discusses of whether there are further institutional veto players and what influence they have will be looked at when we examine the indices of democratic regimes. These three constitutional characteristics coincide with the basic distinctions made by ‘traditional’ institutionalism (Loewenstein 1957). Presidentialism and bicameralism are concerned with the horizontal structure of the democratic system; the former with the relationship between government and parliament, and the latter with the internal structure of parliament. Federalism, in contrast, is concerned with the vertical structure of the democratic system. We can speak of a federal system when there are parliaments and governments with their own powers in territorial units below the national level (states, provinces, Länder). In the political system of a country, these federal entities constitute veto players in policy formation, regardless of whether they are represented at the national level in a upper house. There is extensive and differentiated discussion on the importance and effects of all three constitutional characteristics – presidentialism, bicameralism, and federalism. In the theory of Tsebelis – and in other veto-player approaches (e.g., Immergut 1992) – they are largely left out of account.
Tsebelis (1995) and Tsebelis and Money (1997) relate these characteristics merely to the theoretical criterion of the distribution of power, and with respect to this criterion all three characteristics are functionally equivalent. In the concluding remarks we will address the question whether this degree of abstraction and the functional equivalence of the institutional characteristics thus gained is too great.
5. Indices of Democratic Regimes
The preceding section dealt with two approaches to constructing typologies of democratic regimes, the presidentialism-parliamentarism approach and the veto-player approach. This section describes and discusses the indices that can be assigned to these two approaches.
The two most important indices of the presidentialism-parliamentarism typology are those developed by Shugart/Carey (1992) and Sartori (1994a, 1994b). Both are concerned with only one characteristic of the democratic regime, the relationship (separation or integration) between the executive and the legislative functions. The decisive question for index construction is the operationalization of this relationship. Both indices assume an operational definition of a ‘pure’ presidential system (cf. the preceding section). Sartori (1994b: 106) states: ‘… a system is presidential if, and only if, the head of state (president) (1) receives office by popular election, (2) during his preestablished tenure cannot be discharged by parliamentary vote, and (3) heads the government or governments which he appoints. When all these conditions are met, then we doubtlessly have a ‘pure’ presidential system’ (Shugart 1993: 30 offers a similar definition). On the basis of this definition, the situation initially seems to be clear: ‘Presidential and parliamentary systems are generally defined by mutual exclusion. … To be sure, a presidential system is non-parliamentary, and conversely, a parliamentary system is nonpresidential. However, division of real world cases between these two classes yields both incongruous bed-fellows and dubious inclusions’ (Sartori 1994b: 106). The problem is accordingly to determine the categories that lie between ‘pure’ presidentialism and ‘pure’ parliamentarism.
Shugart/Carey (1992) and Shugart (1993) propose a relatively complex typology that also constitutes an ordinal index: ‘1. ‘Pure’ presidential, 2. President-parliamentary, 3.
Premier-presidential, 4. Parliamentary with ‘president,’ 5. ‘Pure’ parliamentary’ (see table 3). What is particularly important in this typology is the differentiation of presidentialism.
In contrast to ‘pure’ presidentialism, the presidential-parliamentary system has a dualistic executive with president and premier, in which the premier heads the government and is dependent on the president (e.g., Russia). In the premier-presidential system the executive is also dualistic, but the premier is independent of the president (e.g., Poland). Both types of parliamentary system are characterised by a monistic executive. In the case of parliamentarism with a directly elected president there is a president directly elected by the people, but he has either no or very few powers (e.g. Austria).
Shugart/Carey’s typology is notable for the asymmetry between the two poles of pure presidentialism and pure parliamentarism, which is not theoretically accounted for. Basically, it is not so much a presidentialism-parliamentarism typology as a presidentialism typology.
Sartori’s (1994a, 1994b) typology is both more symmetrical and simpler. According to Sartori, the typology of Shugart/Carey is not clear-cut in differentiating the various types and is unnecessarily complex. He restricts himself to the three-way distinction between presidentialism, semi-presidentialism and parliamentarism (see table 3). Which of the two typologies is ultimately the more appropriate can be measured firstly by how well (exhaustively and clearly) the democratic regimes of specific countries can be classified on this basis, and, secondly, by its explanatory power. However, this is not the purpose of our study.
Whereas the presidentialism-parliamentarism typologies concentrate on this one structural characteristic, presidentialism in the veto-player typologies, which take this characteristic into account, is merely one structural characteristic among others. Table 4 shows the selection of structural characteristics by indices of democratic regimes that we ascribe to the veto-player approach. With the sole exception only of Lijphart (1999), the authors Huber et al. (1993), Tsebelis (1995), Colomer (1996), and Schmidt (1996) more or less explicitly acknowledge commitment to this approach.