This is p. 169
two factors, both of which have some "left-right" elements. (See Silverman (1985) for a probing conceptual analysis of two dimensions in the left-right typology.)
In principle, parties can change their ideology, and Downs (1957) theorized that they do change in order to win votes. However, research on the spatial distribution of voters in European party systems disclosed that parties avoid the center of the left-right dimension, even though that area is dense with voters (Listhaug, Macdonald, and Rabinowitz 1990). Despite instances of abrupt changes in party ideology for electoral gains--the German SPD's transformation in 1959 is the stellar example (see Panebianco 1988, 253-257)--the more common argument has been that parties change issue positions incrementally over time, becoming less ideological and more "catch-all" in nature (Kirchheimer 1966). Thomas' longitudinal study of 54 parties in twelve western nations over nearly a century showed "a dramatic narrowing in the scope of domestic political conflict" over issue positions (1975, 46, Thomas 1980). Thomas also uncovered subtle variations of the "end of ideology" and the "convergence" hypotheses, some of which were not supported.
Most recently, scholars have focused on whether the conflict between "materialist" and "post-materialist" values has absorbed the classic left-right conflict in party cleavages, as Inglehart claimed (1977, 242 and 1990, 296-298). Knutsen (1988 and 1989) tested this "new politics" viewpoint with data on party preferences in ten countries and found that both cleavages existed, with neither dominating. In a related vein, Kitschelt (1990) explored the "new structural differentiation and polarity" introduced in Europe by Left-libertarian parties (p.201).
Issue-based parties: While most major parties can be placed along a single left-right continuum, and still more can be comfortably accommodated with the introduction of a second dimension, some parties resist classification because of their commitment to a single issue or type of issue. Lane and Ersson (1991, 273) distinguish between nonstructural issues, which are based on notions of national policies, and structural issues, which are based in social groups and give rise to ethnic, religious, regional, and certain class-based parties (pp.103-111). (Parties based on structural issues are treated below under social support.)
Considerable research has been done recently to identify parties that promote nonstructural issues, especially "new" issues promoted by new parties when existing parties fail to take up the issue (Lawson and Merkl 1988; Dalton and Kuechler 1990). The prime example is protection of the environment advocated by European "Green" parties (Müller-Rommel 1985; Schoonmaker 1988). In their study of 233 new parties formed in western countries from 1960 to 1980, Harmel and Robertson (1985) found that only 10% qualified as "new issue" parties "(i,e., those characterized as ecology parties, anti-nuclear or peace parties, anti-EEC partes, anti-NATO parties, or feminist parties)"--despite the attention academics paid to them (p. 508). Nearly half of the new parties offered alternatives to "old" issues on the left-right dimension. They included "tax protest" parties and other parties on the new right (Lane and Ersson (1991, 108-111). Whether the purer "new politics" parties fit on the Left-Right dimension has been a subject of study (Müller-Rommel 1990). Kitschelt and Hellemans (1990) set forth the issues, and in a limited study of Belgian parties, concluded that economic leftism exists in contemporary politics but it does not exactly match the traditional left-right dimension.
Most research on issue orientation is concerned with measurement. When ideology and issues are involved in theoretical studies, they are mainly treated as independent variables. Scholars are less concerned with the causes of issue orientation than with its consequences for government process (coalition formation) and policy outcomes (e.g, economic and social programs). We review this literature later, for it involves other concepts that we need to discuss first.
Parties are formed not only to promote policy issues in a Burkean fashion but also to secure social interests (Charlot 1989). Lipset and Rokkan (1967) produced the most powerful comparative statement of parties as political expressions of social cleavages. They isolated four major cleavages--center-periphery, state-church, land-industry, and owner-worker--that gave rise to European parties with different social bases, e.g., regional, religious, occupational, and so on. Dalton (1988: 128-149) provided a concise and updated explanation of these cleavages. Although Dix (1989) argued that the Lipset-Rokkan historical cleavages did not apply to Latin American parties, there is evidence that the Lipset-Rokkan social bases do apply across cultures. More specifically, the six social dimensions--economic status, religion, ethnicity (including language and race), region, urbanization, and education--were used productively to analyze party support across a broad range of nations (Janda 1980b, 41).
The operating assumption of an early collection of voting studies was that "social differences structure party loyalties" (Rose 1974, 16). Smith (1989) described the difficulty in keeping straight the distinction between parties and party systems when discussing the subject of