Janda and Gillies: "How well does 'region' explain political party characteristics?" --> back to first page

Compared with many troublesome social science concepts, the concept of region is relatively clear. Vance's definition of region, 'a homogeneous area with physical and cultural characteristics distinct from those of neighboring areas' (Vance and Henderson, 1968: 377) is close to its recent definition by two Soviet scholars as 'a complex spatial socioeconomic system, characterized by a stable combination of political forces and possessing a specific complex of features' (Vitkovsky and Kolossov, 1980: 539). Perhaps the most critical issue in conceptualizing a region revolves around the requirement of geographical contiguity. Must a region be geographically bounded or can it consist of areas that are spatially discontinuous? Russett, who reviews some of the ambiguities and controversies associated with the term, holds that most analysts would reject a simple geographic definition in favor of some criterion of economic and social homogeneity (1967: 2-7). Although Young (1969) criticizes Russett for this broad conceptualization, it is clear that most scholars do not require contiguity in defining a region (Cox, 1969:
71, 77). This is especially true in the international sphere. Thus Vance calls a region 'a group of national states possessing a common culture, common political interests, and often a formal organization' (Vance and Henderson, 1968: 378). Under this conception, the British Commonwealth might qualify as a region, despite its lack of territorial contiguity.
The concept in comparative politics has served to explain similarities in politics and institutions within nations in the same region. In effect, region becomes a synonym for 'culture area', which Ehrich and Henderson define as 'geographical territories in which characteristic culture patterns are recognizable through repeated associations of specific traits. . .' (1968: 563). They find ample evidence for the persistence of culture patterns in given areas over time and even the reappearance of old boundaries for culture areas when peoples with different culture patterns overrun the territory. Although there are pitfalls in 'political culture' explanations of politics, the concept holds some utility for the comparative analysis of political institutions and behavior (Elkins and Simeon, 1979).
The coincidence of cultures with geography invites political scientists to employ regional concepts when they seek to 'hold constant' the effects of political culture on political parties. Thus parties in Western Europe (often including Anglo-American countries) are treated as a group in the writings of Alford (1963), Daalder (1966), Duverger (1961), Epstein (1967), and Merkl (1980). Latin-American parties are treated as a group by Blanksten (1960), Scott (1966), and Ranis (1971). Pye (1960) and Weiner (1960) discuss patterns that are considered common to the parties in Asia, Southeast Asia, and South Asia. Others (e.g. Coleman, 1960; Hodgkin, 1961; Coleman and Rosberg, 1964) find similarities among African parties.

The causal mechanisms in regional explanations

Following a well-established tradition in anthropology, Ross and Homer (1976) cite two causal mechanisms (function and diffusion) to explain similarities of traits in domestic politics among countries. A 'functional' relationship between social and political traits might be based on socialization processes, e.g. authoritarian child-rearing practices producing highiy centralized political parties. Another expression of a functional relationship between traits, seldom discussed by anthropologists, could be game-theoretic: certain institutional arrangements (e.g. proportional representation) produce other institutional responses (e.g. multiple parties). The other causal mechanism for political similarities among nations is 'diffusion', which amounts to simple borrowing of traits or institutions--as in adopting presidential government, bicameralism, proportional representation, or even constitutions. Such borrowing is apt to be especially common among nations in the same region, although communications technology and the high degree of interaction among modem states has led to the diffusion of institutions across the world.
We are less concerned here with identifying the causes of intranational political similarities within regions than with determining the extent of those similarities. Our special interest is in the similarities of party politics among nations in the same region. We believe that the pattern of party politics reflects the central nature of a national political culture as defined by Pye:

A national political culture thus consists of both an elite subculture and a mass subculture, and the relationship between the two is another critical factor determining the performance of the political system. The relationship determines such matters as the basis of legitimacy of government, the freedom and limitations of leadership, the limits of political mobilization, and the possibilities for orderly transfers of power (1968: 220).

Because political parties are designed to link masses and elites, we expect them to be especially responsive to and reflective of national political cultures. To the extent that political cultures are regional as well as national, we expect to find similarities among parties in the same region.

Delineating world regions

Attempts at delineating world regions have produced comparable results (e.g. Russett, 1967; United Nations Secretariat, 1977; Kurian, 1978), with the main difference being in the number of regions identified. Based on demographic characteristics, the United Nations Secretariat divided the world into eight major areas--East Asia, South Asia, Oceania, Latin-America, Africa, Europe, USSR, and Northern America. This is a relatively standard classification, but it is not universally accepted. In the absence of a definitive classification of nations by regions, this regional analysis of party politics begins with the 10 'culture areas' that served as the sampling units for parties studied in the International Comparative Political Parties Project (Janda, 1980). The project identified 92 countries that had functioning party systems from 1957 to 1962. These countries were divided into 10 culture areas:

Anglo-America

West Central Europe

Northern Europe

South America

Central America and the Caribbean

Asia and the Far East

Eastern Europe

Middie East and North Africa

West Africa

Central and East Africa

The first region embraced nations dominated by British political culture; the other nine conformed to common geographic divisions. Five countries were drawn at random from each area, producing a representative sample of party systems within regions and across the world. Party systems were sampled rather than parties to permit study of parties in interaction with other parties. Because this strict sampling procedure did not result in the selection of the United States, the United Kingdom, and Canada, these three countries (desired for substantive interest) were added to the sample and assigned to the Anglo-American cultural area. This small addition of cases to the random sample of party systems was judged to be worth whatever minor bias might be introduced into the analysis.

With the addition of the two North American countries and the United Kingdom, the Anglo-American culture area became the most populated in the sample and the most diverse, given its inclusion of India and the old Rhodesia and Nyasaland Federation. In most political scientists' classifications, all these countries except the latter two belong to the 'Western Community' culture area. We conform to common practice by reassigning India to the Asian category and Rhodesia/Nyasaland to East Africa. We differentiate the Western Community, however, into three subgroups: (a) the six countries remaining in Anglo-America, (b) West Central Europe, and (c) Scandinavia and the 'Benelux' countries of northern Europe. The complete set of 53 countries stratified by region is given in Table 1, which also identifies the parties and reports the data used in the analysis. The concepts heading the columns in Table 1 are discussed below, along with the results of the analysis of variance.