Party change and cleavage formation : the effects of value change on the Spanish party system

Party Change and Cleavage Formation: The Effects of Value Change on the Spanish Party System Mariano TORCAL Universidad Autónoma de Madrid José Ramón MONTERO Instituto Juan March Working Paper n.132 Barcelona 1997 Since the mid-1920s, party systems in Western democracies have developed around the cleavages generated by the modern secular state and industrial society (Lipset and Rokkan 1967). When the initial impact of these forces subsided, Western Europe saw a period of political stability which was reflected in the basic continuity in electoral results from the 1940s until the beginning of the 1970s (Rose and Urwin 1970). The electoral volatility that many scholars have seen since then has, however, challenged this pattern of stability (Maguire 1983; Pedersen 1983; Crewe and Denver 1985)1. The literature on «new politics», for instance, takes the value change hypothesis as the point of departure for the explanation of electoral instability (Dalton et al. 1984). The supporters of this approach maintain that the traditional structure of cleavages which characterized Western politics has lost some of its significance in favour of a new value-based cleavage along the materialist/postmaterialist dimension. As a result, they go on to argue, the traditional left-right schema that lies behind the structure of cleavages has also been affected (Knutsen 1989a; Inglehart 1977, 1984 and 1990a). They also assert that the rise and success of Green parties and the proliferation of alternative electoral lists can be attributed to the increasing prevalence of these new values (Müller-Rommel 1989 and 1990). Spain, like most advanced industrial societies, is undergoing a cultural change (Torcal 1989 and 1992; Scarbrough 1995). Younger generations of Spaniards have substantially different value systems, in which maximum priority is no longer given to physical sustenance and safety, but rather to sentiments of belonging, selfexpression, and quality of life. As in other Western societies, the unprecedented levels of economic prosperity, physical security, and education that have existed over the last few decades are responsible for this shift. How value change has affected Spanish party politics? Much of our interpretation of mass politics in Western Europe is derived from Rokkan’s and Lipset’s freezing hypothesis. Although widely accepted and influential, this approach suffers from two problems. First, it pays scant attention to the underlying values, as opposed to issues, which inform class and religious conflicts (Van Deth and Scarbrough 1995, 43). Second, its treatment of the relationship between social divisions and party politics is too deterministic, leaving little room for human agency, that is, political leadership (Chhibber and Torcal 1997). More recently, Bartolini and Mair (1990, 212-220) have reformulated the concept of cleavage, drawing a distinction between its three constituent elements: social structure, values and beliefs, and institutional and organizational arrangements. A cleavage, therefore, must not only have social and cultural bases, but must also be institutionalized, developing its own «autonomous strength» and acting «as an influence on social, cultural and political life» (1990, 218). In accordance with this definition of a cleavage, Knutsen and Scarbrough (1995) have suggested that value change may be generating value voting at the expense of cleavage voting precisely because of its lack of an institutional or organizational dimension. Independently of the excessively strict interpretation of the «institutional element» of the cleavage, this argument shows that the process of cleavage formation and its consequences for the party system are still seen from the «bottom-up», allowing little room for the role of the elites in the formation of the institutional element of the cleavage. In contrast to this approach, we argue here that elites and party strategies play an important role in the process of cleavage formation. Thus, we maintain that value change is an emerging new cleavage with potentially profound consequences for the party system, but it can only be perceived if we consider two intervening factors: firstly, the generation that individuals belong to; and secondly, the electoral strategies adopted by the elites to capture the vote of postmaterialists. In this paper, therefore, we will consider the extent to which value change is affecting the Spanish cleavage structure. According to the traditional formulation of cleavage theory, value change should lead to some modifications in the party system. This is, for instance, the interpretation defended by Knutsen and Scarbrough (1995). However, we contend that the link between these two levels of a cleavage is mediated by the crafting of the political elites. In Spain, some existing party elites are reacting to the presence of this new value conflict and winning the electoral support of postmaterialists; as a result, no significant change is occurring in the party system, nor are new parties emerging. Cultural change, therefore, is not resulting in party system change, but is rather fuelling party change, i.e., it is altering the character of some existing parties2. Hence, the «autonomous strength» of the institutional dimension of a cleavage depends on political crafting, and its consequences are not as self-evident as the traditional interpretation of cleavages would tend to suggest. Furthermore, the effects of value change on party politics are more noticeable if we take the existence of different political generations into account since, according to Inglehart’s hypothesis, value change is mostly affecting young generations. Thus, its impact on party politics is more conspicuous among these generations than at the aggregate level. We begin by analysing the relationship between cultural change and generational replacement in Spain, showing that replacement is the driving force behind value change. In the rest of the article, we will try to relate the materialist/ postmaterialist dimension to some facets of Spanish politics. We will consider the effects of value change on the structure of political cleavages, examining how value change might be altering the ideological spectrum of some Spanish voters, and conclude with an assessment of the effects cultural change has on the Spanish party system. We will show that the effects of value change on both become more visible if we take the intervening influence of generational replacement and leadership into account. MATERIALISTS AND POSTMATERIALISTS IN SPAIN To what degree are postmaterialist values implanted in Spanish society? To answer this question we have analyzed the data of six surveys carried out in Spain in 1980, 1989, 1991, 1992, 1993, and 19943. From the data for the last fourteen years shown in Table 1, it is evident that the percentage of postmaterialists has increased slightly (with the exception of 1993), while the decline in the percentage of materialists has been remarkable (from 62 per cent in 1980 to 43 per cent in 1993 and 45 per cent in 1994)4. In Spain, therefore, relatively high rates of economic growth did in time lead to relatively high rates of decline in the proportion of materialists, but not a corresponding increase in postmaterialist values, as Inglehart and Abramson (1994) maintain. These affirmations, however, should be made with some caution in the light of the short period of time studied. On the other hand, the proportion of postmaterialists in Spain is very similar to the average for ten European Union countries (excluding Spain and Portugal), confirming that relatively high levels of GNP per capita predict relatively high levels of postmaterialists (Inglehart and Abramson 1994, 346-350). The data included in Table 1, however, also highlight some comparative differences. In Spain, the percentage of materialists is higher than in other EU countries, while the percentage of mixed individuals is lower. It seems, therefore that the relatively high levels of GNP per capita resulting from very rapid economic growth have given rise to a value change that is rather different from that found in other European countries (Scarbrough 1995, 134): a significant proportion of postmaterialists exists alongside the still majoritarian materialist presence5. We argue that these peculiar characteristics of value change in Spain are the result of the particular nature of the profound economic and social transformation experienced in the country since the 1960’s6. The pace of economic growth has produced levels of wealth comparable to other European countries, but also distinct levels of access to this economic prosperity (security) among Spaniards, due to the combination of still notable social inequalities and a welfare state which was essentially constructed during the 1980s. These conditions have given rise to a relatively high proportion of postmaterialists in comparison to other EU countries, but also to the peculiar coexistence of these postmaterialists with a wide plurality of materialists, even though the latter have been declining in relative terms over the last fifteen years. In Spain, as in other advanced industrial societies (Abramson and Inglehart 1986, 1987, 1992 and 1995; Knutsen 1989b; Scarbrough 1995), generational replacement is contributing to the spread of postmaterialism. In this respect, we have used the percentage difference index (PDI) between materialists and postmaterialists by cohorts in 1980, 1989, 1992, and 19937. As can be seen in Figure 1, the younger cohorts account for a large proportion of the postmaterialists, a tendency which is confirmed as each cohort comes of age. Hence the PDIs between younger cohorts are generally lower than those between older cohorts. Figure 1supports the hypothesis of cultural change and generational replacement in Spain, and confirms the results of previous studies (Orizo 1991; Montero and Torcal 1990 and 1992)8. Despite the overall persistence of differences between cohorts, the data also suggest the presence of period effects, as well as some instability in cohort differences, two tendencies which are at odds with the generational replacement hypothesis. The variations in the PDIs shown in Figure 1 can be attributed to three kinds of effects. Those produced by influences associated with aging are age effects; those produced by influences associated with birth cohort membership are cohort effects; and those produced by influences associated with each period of time are period effects9. In order to distinguish between these three effects, we have used a longitudinal research design10. Our conclusion is that, although the data reflect some period effects, the variability of postmaterialists is largely a product of cohort effects, confirming the generational replacement hypothesis. As can be seen in Table 2, different longitudinal models were used to study separately the period, cohort, and age effects on the variations in a postmaterialism index between 1989 and 1993. In the first Model in Table 2, we regressed a postmaterialist index (factor scoring) on twenty dummy independent variables, representing the nine different cohorts, nine age groups, and two periods included in this design11. The coefficients of this partial regression indicate that cohort effects explain most of the variability in the postmaterialist index. Compared with the cohorts effects, the period and age effects were negligible. The sign and magnitude of these coefficients show that the younger the cohort, the higher the postmaterialist index. But none of the coefficients generated by this model are statistically significant, a result which is due to the high multicollinearity between these variables12. After confirming the effect of multicollinearity on the absence of observable statistical significance of cohort effects13, we attempted to resolve this problem by transforming the age variable (quadratic and age minus the average age)14. The influence of cohort effects on the dependent variable, the postmaterialist index, is evident from the results of Model 2 in Table 2, providing further support for the generational replacement hypothesis. The partial regression coefficients of the younger cohorts were indeed both very high and statistically significant. Although the partial regression coefficient of Period 2 is not very important (.126 from 0 to 1), it too is statistically significant, so showing the relative importance, defended by some scholars (Shively 1991; Scarbrough 1995), of period effects in the variability of postmaterialism. We have further confirmed the significance of cohort effects by substituting the nine cohort dummy variables with one variable representing the average education of each cohort15. The results are included in Model 3 in Table 2. The partial regression coefficients of average education by cohort were statistically significant and in the expected direction (in accordance with the cultural change hypothesis). No other regression coefficient was statistically significant. Indeed, cohort replacement seems to be the major force driving value change in Spain. Duch and Taylor (1993) have seen the relative importance of education over GNP per capita on postmaterialism as the principal counter argument against the scarcity hypothesis, which is one of the theoretical cornerstones of Inglehart’s theory16. According to Dutch and Taylor, value change is the result of higher levels of education, rather than different levels of economic prosperity. The Spanish data also seemed to point to the centrality of education. The results of Model 4 in Table 2, which includes the average GNP per capita by cohort, highlights the importance of education as one of the principal factors behind cultural change in Spain: the partial regression coefficient of average education by cohort was the only significant one. In contrast, GNP per capita was not statistically significant. Despite this, however, we disagree with Duch’s and Taylor’s interpretation. As Scarbrough (1995) argues, GNP per capita is only a very imperfect measure of economic security17. This indicator only refers to the overall level of economic prosperity at some point in the past, but does not capture other very important factors shaping economic security, such as the degree to which a nation’s social welfare system has developed. We contend that GNP per capita does not capture economic security in Spain either. As argued above, rapid economic growth in Spain has produced new social inequalities, whilst a fully-fledged welfare system capable of compensating for these is still not in place, resulting in a notable proportion of postmaterialists along with a wide plurality of materialists. In Spain, as in other Western European societies (Scarbrough 1995, 154-6), education better captures the different levels of economic security enjoyed by wide sections of society; this is why education appears in Model 4 in Table 2 as the only significant variable in defining value change. It should also be noted that Spain has the highest unemployment rate among Western countries18, only confirming that rapid economic growth has not benefited all sectors of Spanish society equally. In this respect, Clarke and Dutt (1991) have argued that the rising levels of unemployment during the 1980s have been producing, contrary to what the cultural change hypothesis suggests, an increase in postmaterialism. Recently, Inglehart and Abramson (1994 and 1995) have demonstrated that unemployment has only fuelled a temporary increase in the number of materialists (period effects). The Spanish case may contribute to this debate by showing that consistently high levels of unemployment together with rapid economic growth have produced coexistence of significant number of postmaterialists with a majority of materialists. Abramson’s and Inglehart’s (1993) argument that the effects of economic growth are different from the effects of economic security in terms of value change are therefore correct. Economic growth alone cannot produce fully-fledged value change. For this to take place, it must provide economic security for the majority of the society and should be accompanied by high levels of education. The distinctive features of value change in Spain are the result of a combination of rising levels of education and the economic security enjoyed by some sectors of Spanish society. THE STRUCTURE OF POLITICAL CLEAVAGES AND VALUE CHANGE IN SPAIN In most Western Europe countries, the left-right schema summarizes the universe of political conflicts and issues for people, making these more understandable, meaningful, and manageable. More importantly, the meaning of this schema is defined by the basic structure of conflicts in a given social system (Fuchs and Klingemann 1990, 223). In Spain too, and in contrast to party identification19, the images of left and right are firmly anchored, and reflect the basic structure of political and social conflicts. The classical cleavages of Spanish society are mostly represented on this ideological continuum, they have a major influence on political attitudes and behaviour, and constitute the most important frame of reference for evaluations of political parties, political leaders, and some political institutions (Linz et al. 1981; Sani and Montero 1986). Spaniards’ self-placement on the left-right scale has been very stable over time (Montero 1994). But, as other scholars have noticed, the apparent underlying stability of the scale may conceal an important change in its meaning (Fuchs and Klingemann 1990). How important is the materialist/postmaterialist conflict in defining the current meaning of the left-right continuum in Spain? To answer this question, we have adopted two of the main approaches to the study of political cleavages (Knutsen 1989a, 495). One approach has used statistical techniques such as regression analysis to identify the different social characteristics that may explain political preferences and behaviour. The other has sought to identify empirical cleavages using dimensional analysis. In the next two sections we will present the results obtained when these two approaches were used to examine the effect of value change on the structure of political cleavages in Spain. Postmaterialism and the Structure of Political Cleavages In order to assess the importance of the materialist/postmaterialist conflict in defining the meaning of the ideological continuum, we regressed a postmaterialist index (factor scores) on the left-right scale. We included in the model two independent variables representing traditional cleavages in Spanish society: materialist class conflict20 and the religious cleavage21. We also added income level to the model in order to provide an objective measure of class. Finally, since gender has been shown to have some effects on the ideology of Spaniards (Díez Medrano et al. 1989), this variable was also incorporated. Analysis of the 1993 data produces some interesting results. Table 3 immediately suggests that class-consciousness and religious cleavages are still remarkably important for the social definition of the left-right schema. Compared with the impact of these variables, the postmaterialist and gender coefficients appeared to be irrelevant. Do these results provide definitive evidence that value change has had no significant impact on the left-right continuum? In order to answer this question we must consider the distinct distribution of materialists and postmaterialists in the different generations of Spaniards. As we have noted above, rapid economic growth and social modernization in Spain has given rise to a situation in which a relatively large number of postmaterialists coexist with the plurality-holding materialists. Therefore, it could be expected that some traditional cleavages would remain the dominant force defining the general structure of political conflicts. But, as we have also seen here, the proportion of postmaterialists is much higher among younger cohorts, and it is precisely these generations which constitute the main source of changes in the structure of political cleavages (Knutsen 1995, 181-182). Cohort differences, therefore, must be introduced in order to ascertain the extent to which the generational factor affects both the structure of political cleavages and their impact on the left-right continuum. Table 4 confirms our expectations. We repeated the regression model performed in Table 3 with the 1993 data, but this time including in the model a new set of variables which represent the interaction of political cleavages and cohort effects22. The resulting picture shows that subjective social class and religiosity continue to appear significant despite the generation effects. But it also shows the importance of the materialist/postmaterialist dimension in defining the left-right schema when cohort interaction effects are specifically considered. We contend that these results, together with the generational replacement observed in the previous section, clearly indicate that traditional materialist cleavages are giving way to new values in the constellation of meanings which define political conflicts. Of the traditional cleavages, religiosity still exerts a dominant influence, but it becomes significant only among older generations, confirming the depolarizing effects of secularization in Spanish politics. Class conflict also remains important. Therefore, it seems that in Spain, as in many other Western European countries (Knutsen 1988; Inglehart 1990b and 1990c, Van Deth and Scarbrough 1995), more value-oriented or ideological cleavages, including new cultural and religious values, are increasingly those which shape the left-right schema. The changing meaning of the left-right schema and the increasing importance of the materialist/postmaterialist conflict may be seen more clearly by examining the relationship between the scale and various issues. A postelectoral survey carried out in 1993 included ten questions referring to a selection of socially and politically significant issue attitudes23. Respondents were asked to place themselves on a 1 to 10 scale according to their opinion on two contrasting statements related to basic issues24. A factor analysis of these ten scales (not shown here) revealed the existence of four different dimensions. The first is the religious/secular dimension, which is defined by the abortion issue and the authoritarian/libertarian conflict25. The second dimension points to the conflict between increasing social services and reducing taxes, and also includes the issue of European economic integration. The third refers to the debate about the privatization of public corporations and competition in the workplace. Finally, the materialist/postmaterialist dimension encompasses the conflicting issues of the competition between them on the postmaterialist dimension. Nevertheless, until now this competition has proved extremely one-sided. IU has successfully managed to transform itself from a typical communist party of the old left into the main electoral option advancing the characteristic postmaterialist issues of the new left. Thanks to this change, the Green parties appear to have lost their electoral battle with IU. The moment of greatest support for the Green parties came in the 1989 general elections, when the six different electoral lists won 1.7 per cent of the vote, and in the European elections held that year, when four Green lists obtained 2.9 per cent of the vote, although in neither case did they win any seats (Baras 1992). Since then, however, the relative growth of IU has contrasted with the virtual disappearance of the Green parties: they won only 0.86 of the vote in the 1993 general elections, and 0.28 in 1996. Part of IU’s success may be attributed to the errors made by the Green parties themselves. The Green parties in Spain are among the most divided, and weakest, of all their Western European counterparts. Their internal history of fragmentation and splits in a context of electoral failures has reproduced the tendency seen in some other Western Green parties. Their external strategy illustrates the hypothesis that factors other than a pure logic of electoral competition sometimes shape the dynamics of parties (Kitschelt 1989). A second effect of value change on Spanish politics can be seen in the heterogeneity of IU’s voters. Their very distinct profiles not only reflect the habitual heterogeneity resulting from catch-all strategies, but above all the differences between their voter’s basic preferences. For IU, value change has created a potential basis for intra-party conflict between its blue-collar, staunchly communist voters, its traditionally leftist lower-middle class electorate, and its younger upper-middle class postmaterialists. In any event, the case of IU is rather unique, even in Southern Europe (Bosco 1995): the PCE, through Izquierda Unida, is the only Communist party which, whilst remaining so, has successfully been able to channel a significant degree of support of postmaterialist voters; support which until now has condemned the Green parties to being completely marginal groups within the party system. In short, cultural change in Spain is producing changes in a party rather than changes in the party system as a whole; that is, value change is producing party change rather than party system change. Moreover, and in the third place, value change in Spain is redefining to some extent the spatial competition between parties on the left. In striking contrast to many European party systems, in which Green parties compete electorally with larger socialist parties (Müller-Rommel 1989 and 1994), competition in Spain continues to be between IU and the PSOE. But this competition is open, and still far from crystallized. In the early 1990s, the Socialist Party’s long period in power explained the relatively scant presence of postmaterialists among its electorate: only 19 per cent of PSOE voters, compared with 39 per cent of those of IU, were postmaterialists in 1989 (Montero and Torcal 1992). It will be remembered that Figure 5 provided eloquent graphic confirmation of the position occupied by the PSOE. However, with its return to the opposition in 1996 the competition between the two main parties on the left has restarted. The outcome of this competition will hinge on how the major parties address new issues emerging from value change, at the same as they deal with the demands of their traditional electorates. More particularly, it will depend on how successful IU is in processing simultaneously the demands of its potentially conflicting groups of voters. It will also depend on how successful the Socialist Party is in modernizing and renewing its electoral appeal from its place in the parliamentary opposition, and hence in competing with IU for postmaterialist voters on new grounds. Finally, it will also be influenced by the extent to which the Spanish Green parties are able to overcome their “natural tendency” (Kistchelt 1990) to division, and so move, if only a little, up the league table of electoral support for Green parties in Europe. In this way, as in many other Western European party systems, the distinct effects of value change on party politics are conditioned by several political and institutional factors, as well as by the different strategies adopted by political leaders40. The effects of value change, combined with these intervening factors, may lead to either party change or party system change, or both. In Spain, value change is only producing the former, and is mainly restricted to the established left-wing parties, and above all to the Communist-dominated coalition, IU. CONCLUSIONS In this paper we have examined in some detail the formation of a new cleavage as an empirical illustration of its impact on party politics. We began by analyzing the magnitude of value change in Spain, and how recent economic, social and political events contribute to a better understanding of the particular character of change. The last three decades have seen enormous economic growth, social modernization, and political change in Spain. Rapid economic growth has generated comparable levels of wealth to most Western societies, but also different levels of general economic security. The persistence of some major social inequalities, combined with a still comparatively limited welfare state, has created distinct levels of perception of economic goals and priorities among Spaniards. This has produced a relatively large number of postmaterialists who coexist with a majority of materialists. As in many other Western countries, in Spain too generational replacement or cohort effects seem to be one of the main driving forces behind cultural change. Younger «new citizens» are slowly but surely replacing older materialist generations. Postmaterialism, therefore, is relatively widespread among younger cohorts, and it is precisely these citizens who are the source of changes in the structure of political cleavages. As they enter the political arena, they are also the potential source of change in the ideological definitions of parties which embrace new values, and consequently of the party system. Thus, they play an extremely important role in the formation of the new cleavage which results from value change. In this paper we hope to have demonstrated that the political consequences of value change can only be perceived if we take into account generational replacement and the corresponding cohort effects. But Spain also provides a useful opportunity to analyse the impact of value change on the party system. In strong contrast to the axiomatic assumptions frequently made with respect to the relationships between the three elements of a cleavage (a social division, a common set of values, and an organization which institutionalizes it), the Spanish case highlights the levels of autonomy existing in the organizational element. The leaders of the main left-wing parties were able to prevent the emergence of new parties which expressed the political priorities of these «new citizens». The leaders of the existing parties managed to attract these postmaterialist voters and incorporate the postmaterialist dimension into the established framework of electoral competition, whilst the Green parties, paralysed by their leaders’ strategies of division and fragmentation, have been unable to enter the party system. It is clear that the (few) voters of these new parties are mostly postmaterialists, regardless of the cohort they belong to; but these organizations have only obtained remarkably poor electoral results. Their competitors on the left, IU and PSOE, have managed to capture the postmaterialist vote, but they have also experienced a number of changes as result of values change. Generational replacement is having particularly important consequences for the composition of the electorate of IU, the coalition which has proved most successful in capturing the support of «new citizens». This process is increasing the heterogeneity of IU, and hence, the potential for intra-party conflict between the younger and older generations, as well as between its sometimes rather incompatible policy aims. Thus, the effects of value change are largely fuelling a process of party change, rather than party system change. Can this trend be expected to continue as the number of postmaterialists rises in Spain? The change observed, unusual as it may be if one remembers that the successful competitor of the Green parties is an electoral coalition dominated by the Spanish Communist party, has not modified the position of IU within the party system. The PSOE still enjoys a large number of competitive advantages with respect to IU -in terms of its leaders, voters, resources, and institutional presence. Despite the significant proportion of postmaterialists in Spanish society, and the intensity of generational replacement in value change, the strategies developed by the leaders of IU and the PSOE have reinforced the relative autonomy of the party elites insofar as they have been able to prevent the emergence of an autonomous organization from the postmaterialist cleavage. Up to now at least, this margin of manoeuvre for interparty competition has been translated into some changes within the parties, but the party system itself has not been modified by the entry of new parties or more significant changes in any of the existing parties. Table1 Table2 Table3 Table4 Table5 Table6 Figure5 NOTES Two previous drafts of this paper were presented at the ECPR Workshop on «Quantitative and Comparative Research on Values and Ideology» (Oslo, March-April, 1996), and at a seminar held at the Institut de Ciències Polítiques i Socials (Barcelona, February, 1997). Gabriel Colomé, Oddbjorn Knutsen, Wolfgang Jagodzinski, Ferran Requejo, and Elinor Scarbrough made very insightful comments. We also want to thank Justin Byrne, Pradeep Chhibber, Scott Menard, Brian Pollins and Shannon Sullivan for their invaluable help, and to acknowledge the financial support of the Spanish Comisión Interministerial de Ciencia y Tecnología (CICYT) and the facilities provided by the Instituto Juan March. Additionally, M. Torcal would like to express his gratitude to the Department of Political Science at the University of Michigan. 1.The electoral instability of the 1970s has been disputed by some scholars. Ersson and Lane (1982) and Shamir (1984), for example, have found that the aggregate electoral results since the beginning of the century show remarkable patterns of stability as well as instability. Nevertheless, in a more comprehensive study, Bartolini and Mair (1990) have supported the stability hypothesis with data from the period since 1885. 2.Whereas party change alludes to changes occurring in one or various parties in a given system, party system change refers to changes taking place in the party system itself (Mair 1989 and 1990; Bartolini and Mair 1990). For other definitions of party change and party system change see Laver (1989), Smith (1989), and Harmel and Janda (1994). 3. The 1980 and 1989 surveys were conducted by the Centro de Investigaciones Sociológicas (CIS). The 1980 survey was undertaken in November, and 1989 survey in January. The surveys for 1991, 1992, 1993 and 1994 were carried out in June of those years by the Centro de Investigaciones sobre la Realidad Social (CIRES). 4.In Table 1 we have defined individuals in accordance with their factor scores on the materialist/postmaterialist dimension (first principal component); we have used the regression method to create and save the factor scores. 5.Studies carried out in 1990 (Orizo 1991, 43-46) confirm the constant increase in postmaterialists (20 percent) and mixed individuals (58 percent), and the consequent decline in materialists (22 percent). This distribution seems to dilute the peculiar importance of materialists in Spain, apparently indicating that the significant increase in the proportion of mixed individuals has taken place at the expense of the materialists. Although the results of Orizo’s study also confirm the cultural change hypothesis, he used a rather different method to ours when assigning respondents to the categories of materialist, mixed, and postmaterialist (Montero and Torcal 1992, 71). 6.The Spanish economy grew by an average of 7.3 per cent per year in real terms between 1961 and 1970. This growth led to the profound modernization of Spanish society. In 1960, for example, 41.5 per cent of the active population was employed in the primary, 23.3 per cent in the secondary, and 28.1 per cent in the tertiary sector. In 1991, the distribution of the active population was 10.3, 21.3, and 51.9 per cent respectively (see Cuadrado Roura 1990). For comparative data on economic growth for this period, see Van Deth (1995, 51-60). 7.A cohort is defined as «those people within a geographically or otherwise delineated population who experienced the same significant life event within a given period of time» (Glenn 1977, 8). Cohort boundaries can be arbitrarily defined in terms of «any standard period of time», or can be based on some historical events. In this study we have set the cohort boundaries in reference to a number of historical, social, and economic events in recent Spanish history. For a more detailed explanation, see Torcal (1989). 8.Torcal (1992) has stressed the significance of cohorts for the definition of the materialist/postmaterialist dimension, and has demonstrated that value change in Spain follows a clear time developmental model or circular dimensionality, in which the younger the cohort, the closer its position to the most postmaterialist and libertarian items. For the concept of circular dimensionality, see also Weisberg (1974). 9.Unfortunately, a linear relationship between cohort, period and age effects leads each to be confused with the other. Thus, these effects cannot be clearly distinguished from Figure 1 alone. The influence of these three effects is difficult to discern for two additional reasons. As noted by Glenn (1977, 12), first, because in survey data «much of the variation in percentages (or other values) among cells of a cohort table reflects sampling variability», and second, because «as a birth cohort grows older, it suffers attrition due to the death of some of its members». 10. Some authors use the adjective longitudinal to refer to both cohort trend studies with repeated crosssectional studies and panel studies (Glenn 1977; Menard 1991), whereas other terminological purists only apply this adjective to panel studies (Campbell and Stanley 1963). 11. These dummy variables were created from a cohort table using only two years (1989 and 1993). The intervals of age and cohort have to correspond exactly with the interval of time between the surveys. Therefore, age and cohort were grouped in five-year intervals, which correspond to the five years between these two surveys (1989-1993). We obtained nine five-year intervals of cohort and age, excluding, for theoretical reasons, those individuals over sixty. This is also the reason why we could not include more periods in the model. The interval of time between the other surveys was different, making it impossible to include them in the model. For a detailed discussion of this model, see Mason et al. (1973). 12. Tolerance among all these variables was very low, and the variance-inflation factor (VIF) very high, indicating high collinearity among the variables. Linear relationships between the independent variables make the least-squares regression coefficients unstable. Coefficient standard errors are inflated, leading to the possibility that the null hypothesis might not be rejected when it is in fact not true (error type II). Small changes in the data may substantially modify the regression coefficients, whilst scarcely raising the sum of squared residuals (SME). See Fox (1991, 10-21). 13. To confirm the statistical significance of some of the cohort variables, we performed a Joint Hypothesis Testing for Collinearity, finding a significant F statistic. Multicollinearity among these variables was, therefore, causing us to fail to reject the null hypothesis. 14. These are standard procedures in many longitudinal studies attempting to evaluate period, cohort, and age effects. For a detailed discussion of these questions, including some applications, see Menard (1992). 15. Graaf and Evans (1996, 612) state that the «advantages of so doing are twofold. First, the identification problem can be solved, and second the predictions become more informative because the characteristics that should impact on a given dependent variable are clearly specified». 16. Graaf and Evans (1996, 627) also maintain that the severity of war experiences and education are the most important predictors of postmaterial values. 17. In a more detailed analysis, Graaf and Evans (1996, 627) reach the same conclusion, although the preferential relation between education and postmaterialism could have more implications than those considered by Inglehart (1990a and 1995). 18. The Spanish unemployment rate was 5.7 per cent in 1977, reached a maximum of 24.1 per cent in 1994, and stood at 22.4 per cent in 1996. 19. Party identification in Spain is only a secondary variable for the analysis of political conflicts and electoral behavior. Regardless of the measure used, the levels of party identification observed among Spaniards are extremely low (Gunther and Montero 1994). 20. We had to use the subjective dimension of class, as reflected in survey answers to the question «What social class do you think you belong to?» We would have like to have had some other variables to measure objective social class, but unfortunately they were not included in this survey. Income was, however, included in the CIRES surveys, but it was not found to be statistically significant. 21. The variable used here is the frequency of church attendance. Because of the small number of cases, we have not included any variable measuring regional conflicts, which are obviously extremely relevant in Catalonia and the Basque Country. 22. We grouped the six cohorts of Figure 1 into three cohorts, creating dichotomous variables with each. We multiplied the two youngest and oldest cohorts by each variable measuring some political cleavage. Finally, we regressed the left-right scale on those new variables representing the interaction of political cleavages with older and younger cohorts. 23. The survey, codirected by Richard Gunther, Francisco Llera, José María Maravall, José Ramón Montero, and Francesc Pallarès, was carried out by DATA, and financed by the CICYT (SEC95-1007), as part of the Comparative National Election Project (CNEP). 24. The ten issue scales, ranging from 1 to 10, were as follows: (1) «Defend our traditional religious and moral values» vs. «Defend the individual right to be and belief whatever he or she wants»; (2) «Protect the environment to make our cities and countryside more habitable and beautiful» vs. «Promote economic growth without environmental restrictions on corporations»; (3) «Improve social and public services» vs. «Cut taxes»; (4) «Distribute income more evenly» vs. «Create more incentives for individual improvement»; (5) “Nurture the European Union» vs. «Maintain Spain’s national independence»; (6) «Maintain order» vs. «Defend individual rights»; (7) «Privatize public corporations» vs. «Maintain the existing public corporations»; (8) «Declare abortion illegal» vs. «Declare abortion legal»; (9) «Work hard and compete for individual improvement» vs. «Maintain solidarity, and not competition, with workmates and a good atmosphere in the workplace»; and (10) «Increase citizens’ participation in important governmental decisions» vs. «Allow the government to make rapid decisions based on its expertise». 25. For a detailed analysis of the strength of this authoritarian/libertarian dimension as a component of the materialist/postmaterialist conflict see Torcal (1992). 26. The regression obtained was the following: _=4.6460.983X1 + 0.270X2 - 0.564X3 + 0.000X4. All the coefficients were statically significant for p< 0.05 with the exception of b4. 27. It would have been better to treat these interactions as dummy variables, but in this 1993 survey age has already been grouped into seven categories, making it impossible to repeat the cohort analysis performed above 28. The different positions on the scale were measured through dichotomous variables. 29. This fact explains the position of issues, such as «fight social inequalities», close to the postmaterialist items. For a detailed discussion of the interactions between the two conflicts see Torcal (1992). For the concept of sociotropic interest, see Kinder and Kiewiet (1979). 30. In the 1993 survey a 1-7 ideological scale was used. In this way we grouped these values as follows: (position 1) extreme-left; (2 and 3) left and centre-left; (4) centre; (5 and 6) right and centre-right; and (7) extreme-right. 31. In the four general elections held between 1982 and 1993, the Socialist Party won between of 48.4 and 38.8 per cent of the vote (between 58 and 45 per cent of the seats), which allowed it to form single-party governments. Support for the conservative party rose from 26 per cent of the vote in the eighties to 34.8 per cent (40 per cent of the seats) in 1993. In the 1996 elections, the PP won power with 38.8 per cent of the vote (45 per cent of the seats), whilst the Socialist Party returned to the opposition with 37.5 per cent of the vote (40 per cent of the seats). 32. During the eighties, the Communist Party (Partido Comunista de España, PCE) and later IU won between 4 and 9 per cent of the vote. In the 1996 elections IU obtained 10.6 per cent of the vote and 6 per cent of parliamentary seats. 33. In the last elections, their combined share of votes and seats was some 6 percent, but their strategic position has enabled them to play a decisive role in supporting the minority governments of the PSOE and then the PP. 34. We had to exclude the regional (and most significantly the Catalan and Basque) parties from this multidimensional analysis due to the small number of cases, and in spite of the prominent role they play within the Spanish party system. 35. In the Spanish case, the success of the traditional parties has been favoured by the persistent organizational and electoral failure of the Green parties. Since the mid-eighties, their leaders have proved incapable of implementing strategies of fusion of the numerous groups existing, of reducing the excessive number of lists presented in elections, or avoiding highly divisive ideological disputes. See Baras (1992). 36. Klingemann, Hofferbert, Budge et al. (1994, 272) use sentences as quantitative units for their analyses of election manifestos. Although their system seems more reliable, for the purpose of this paper we have relied in the simpler technique of counting lines rather than sentences. 37. We should also mention that IU has adopted different names in some Spanish regions, some of which refer to the Verdes (Greens). For instance, according to the Fourth Federal Assembly hold in 1994, IU’s official name in Andalusia is Izquierda Unida Los Verdes-Convocatoria por Andalucía; and in Catalonia, Iniciativa per Catalunya-Els Verds. 38. The postelectoral 1993 survey reflects some short-term effects produced by the election campaign. For this reason, we have opted to use the 1992 survey in this part of the paper. 39. It should be noted that the Green parties were not included in the questionnaire in 1989, and therefore we could not take them into account in Figure 5. 40. For a more detailed discussion of the sources of change and constraints in party politics, see Rohrchneider (1993). REFERENCES ABRAMSON, Paul R. and INGLEHART, Ronald: «Generational Replacement and Value Change in Six West European Societies», American Journal of Political Science 30/1986, p. 1-25. ABRAMSON, Paul R. and INGLEHART, Ronald: «Generational Replacement and the Future of PostMaterialist Values», The Journal of Politics 49/1987, p. 231-241. ABRAMSON, Paul R. and INGLEHART, Ronald: «Generational Replacement and Value Change in Eight West European Societies», British Journal of Political Science 22/1992, p. 183-228. ABRAMSON, Paul R. and INGLEHART, Ronald: «Education, Security and Postmaterialism», Manuscript, 1993. ABRAMSON, Paul R. and INGLEHART, Ronald: Value Change in Global Perspective. Ann Arbor, University of Michigan Press, 1995. BARAS, Monserrat: Los partidos políticos verdes en Cataluña. Barcelona, Institut de Ciències Polítiques i Socials, Working Paper 47/1992. BARTOLINI, Stefano and MAIR, Peter: Identity, Competition and Electoral Availability. The Stabilization of European Electorates 1885-1985. Cambridge, Cambridge University Press, 1990. BOSCO, Anna: «Four Actors in Search of a Role: The South European Communist Parties», Manuscript, 1995. CAMPBELL, Donald T. and STANLEY, Julian C.: Experimental and Quasi-Experimental Designs for Research. Boston, Houghton Mifflin, 1963. CLARKE, Harold D. and DUTT, Nitish: «Measuring Value Change in Western Industrialized Societies: The Impact of Unemployment», American Political Science Review 85/1991, p. 905-920. CHHIBBER, Pradeep and TORCAL, Mariano: «Elite Strategy, Social Cleavages, and Party Systems in a New Democracy: Spain», Comparative Political Studies 30/1997, p. 27-53. CREWE, Ivor, and DENVER, David (eds.): Electoral Change in Western Democracies. Patterns and Sources of Electoral Volatility. New York, St. Martin’s Press, 1985. CUADRADO ROURA, Juan R.: «Cambio estructural, terciarización y remodelación territorial» in GARCÍA DELGADO, José L. (ed.): Economía española de la transición y la democracia. Madrid, Centro de Investigaciones Sociológicas, 1990. DALTON, Russell J., FLANAGAN, Scott P. and ALLEN BECK, Paul (eds.): Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment? Princenton, Princenton University Press, 1984. DÍEZ MEDRANO, Juan, GARCÍA-MON, Blanca and DÍEZ NICOLÁS, Juan: «El significado de ser de izquierdas en la España actual», Revista Española de Investigaciones Sociológicas 45/1989, p. 9-41. DUCH, Raymond M. and TAYLOR, Michael: «Postmaterialism and the Economic Condition», American Journal of Political Science 37/1993, p. 446-479. ERSSON, Svante and LANE, Jan-Erik: «Democratic Party Systems in Europe: Dimensions, Change and Stability», Scandinavian Political Studies 5/1982, p. 67-96. ERSSON, Svante: «Value Changes in Industrial Societies», American Political Science Review 81/1987, p. 1303-1319. FOX, John: Regression Diagnostics. Beverly Hills, Sage Publications, 1991. FUCHS, Dieter and KLINGEMANN, Hans-Dieter: «The Left-Right Schema» in JENNINGS, M. Kent; VAN DETH, Jan W. et al: Continuities in Political Action. A Longitudinal Study of Political Orientations in Three Western Democracies. Berlin, Walter de Gruyter, 1990. GLENN, Norval D.: Cohort Analysis. Beverly Hills, Sage Publications, 1977. GRAAF, Nan D. de, and EVANS, Geoffrey: «Why Are the Young More Postmaterialist? A Cross-National Analysis of Individual and Contextual Influences on Postmaterial Values», Comparative Political Studies 28/1996, p. 608-635. GUNTHER, Richard and MONTERO, José Ramón: «Los anclajes del partidismo: un análisis comparado del comportamiento electoral en cuatro democracias del sur de Europa» in DEL CASTILLO, Pilar (ed.): Comportamiento político y electoral. Madrid, Centro de Investigaciones Sociológicas, 1994. HARMEL, Robert and JANDA, Kenneth: «An Integrated Theory of Party Goals», Journal of Theoretical Politics 6/1994, p. 259-287. INGLEHART, Ronald: The Silent Revolution: Changing Values and Political Styles among Western Publics. Princenton, Princenton University Press, 1977. INGLEHART, Ronald: «The Changing Structure of Political Cleavages in Western Society» in DALTON, Russell J., FLANAGAN, Scott P. and ALLEN BECK, Paul (eds.): Electoral Change in Advanced Industrial Democracies: Realignemt or Dealignment? Princenton, Princenton University Press, 1984. INGLEHART, Ronald: Culture Shift in Advanced Industrial Society. Princenton, Princenton University Press, 1990a. INGLEHART, Ronald: «From Class-Based to Value-Based Politics» in MAIR, Peter (ed.): The West European Party System. New York, Oxford University Press, 1990b. INGLEHART, Ronald: «Political Value Orientations» in JENNINGS, M. Kent, VAN DETH, Jan W. et al.: Continuities in Political Action. A Longitudinal Study of Political Orientations in Three Western Democracies. Berlin, Walter de Gruyter, 1990c. INGLEHART, Ronald and ABRAMSON, Paul A.: «Economic Security and Value Change», American Journal of Political Science 88/1994, p. 336-354. KINDER, Donald and KIEWIET, Ronald: «Sociotropic Politics: The American Case», British Journal of Political Science 11/1979, p. 129-161. KITSCHELT, Herbert: The Logic of Party Formation. Ecological Politics in Belgium and West Germany. Ithaca, Cornell University Press, 1989. KITSCHELT, Herbert: «New Social Movements and the Decline of Party Organization» in DALTON, Russell J. and KUECHLER, Manfred (ed.): Challenging the Political Order. New Social and Political Movements in Western Democracies. New York, Polity Press, 1990. KLINGEMANN, Hans-Dieter, HOFFERBERT, Richard I., BUDGE, Ian et al.: Parties, Policies, and Democracy. Boulder, Westview Press, 1994. KNUTSEN, Oddbjorn: «The Impact of Structural and Ideological Party Cleavages in West European Democracies: A Comparative Empirical Analysis», British Journal of Political Science 18/1988, p. 323-352. KNUTSEN, Oddbjorn: «Cleavage Dimensions in Ten West European Countries. A Comparative Empirical Analysis», Comparative Political Studies 21/1989a, p. 495-534. KNUTSEN, Oddbjorn: «The Priorities of Materialist and Post-Materialist Political Values in the Nordic Countries - A Five-nation Comparison», Scandinavian Political Studies 12/1989b, p. 221-243. KNUTSEN, Oddbjorn: «Left-Right Materialist Value Orientations» in VAN DETH, Jan W. and SCARBROUGH, Elinor (eds.): The Impact of Values. Oxford, Oxford University Press, 1995. KNUTSEN, Oddbjorn and SCARBROUGH Elinor: «Cleavage Politics» in VAN DETH, Jan W. and SCARBROUGH, Elinor (eds.): The Impact of Values. Oxford, Oxford University Press, 1995. LAVER, Michael: «Party Competition and Party System Change. The Interaction of Coalition Bargaining and Electoral Competition», Journal of Theoretical Politics 1/1989, p. 301-324. LINZ, Juan J., GÓMEZ REINO, Manuel, ORIZO, Francisco A. and VILA, Darío: Informe sociológico sobre el cambio político en España, 1975-1981. Madrid, Euramérica, 1981. LIPSET, Seymour M. and ROKKAN, Stein: «Cleavages Structures, Party Systems, and Voters Alignments: An Introduction» in LIPSET, S.M. and ROKKAN, S. (eds.): Party Systems and Voter Alignments: Cross-National Perspectives. New York, Free Press, 1967. MAGUIRE, Maria: «Is There Still Persistence? Electoral Change in Western Europe, 1948-1979» in DAALDER, Hans and MAIR, Peter (eds.): Western European Party Systems. Continuity and Change. Beverly Hills, Sage, 1983. MAIR, Peter: «The Problem of Party System Change», Journal of Theoretical Politics 1/1989, p. 251-276. MAIR, Peter: «Parameters of change» in MAIR, P. (ed.): The West European Party System. New York, Oxford University Press, 1990. MASON, Karen O., MASON, William M., WINSBOROUGH, H.H. and POOLE, W. Kenneth: «Some Methodological Issues in Cohort Analysis of Archival Data», American Sociological Review 38/1973, p. 242258. MENARD, Scott: Longitudinal Research. Beverly Hills, Sage Publications, 1991. MENARD, Scott: «Demographic and theoretical Variables in the Age-Period-Cohort Analysis of Illegal Behaviour», Journal Research Crime and Delinquency 29/1992, p. 129-199. MONTERO, José Ramón: «Sobre las preferencias electorales en España: fragmentación y polarización (1976-1993)» in DEL CASTILLO, Pilar (ed.): Comportamiento político y electoral. Madrid, Centro de Investigaciones Sociológicas, 1994. MONTERO, José Ramón: «Secularization and Cleavage Decline: Religiosity, Electoral Behaviour and Generational Change in Spain». Paper presented at the ECPR Joint Session of Workshops, Bern (Switzerland), 1997. MONTERO, José R. and TORCAL, Mariano: «Voters and Citizens in a New Democracy: Some Trend Data on Political Attitudes in Spain», International Journal of Public Opinion Research 2/1990, p. 116-140. MONTERO, José R. and TORCAL, Mariano: «Política y cambio cultural en España: una nota sobre la dimensión postmaterialista», Revista Internacional de Sociología 1/1992, p. 61-99. MONTERO, José R. and TORCAL, Mariano: Value Change, Generational Replacement and Politics in Spain. Madrid, Instituto Juan March, Working Paper 56/1994. MÜLLER-ROMMEL, Ferdinand (ed.): New Politics in Western Europe. Boulder, Westview Press, 1989. MÜLLER-ROMMEL, Ferdinand: «New Political Movements and ‘New Politics’ Parties in Western Europe» in DALTON, Russell J. and KUECHLER, Manfred (eds.): Challenging the Political Order. New Social and Political Movements in Western Democracies. New York, Polity Press, 1990. MÜLLER-ROMMEL, Ferdinand: Green Parties under Comparative Perspective. Barcelona, Institut de Ciències Polítiques i Socials, Working Paper 99/1994. ORIZO, Francisco A.: Los nuevos valores de los españoles. España en la Encuesta Europea de Valores. Madrid, Fundación Santa María, 1991. PEDERSEN, Mogens N.: «Changing Patterns of Electoral Volatility in European Party Systems, 1948-1977: Explorations in Explanation» in DAALDER, Hans and MAIR, Peter (eds.): Western European Party Systems. Continuity and Change. Beverly Hills, Sage, 1983. ROHRSCHNEIDER, Robert: «Environmental Belief Systems in Western Europe. A Hierarchical Model of Constraint», Comparative Political Studies 26/1993, p. 3-29. ROSE, Richard, and URWIN, Derek W.: «Persistence and Change in Western Party Systems Since 1945», Political Studies 18/1970, p. 287-319. SANI, Giacomo, and MONTERO, José R.: «El espectro político: izquierda, derecha y centro» in LINZ, Juan J. and MONTERO, José R. (eds.): Crisis y cambio: electores y partidos en la España de los años ochenta. Madrid, Centro de Estudios Constitucionales, 1986. SCARBROUGH, Elinor: «Materialist-Postmaterialist Value Orientations» in VAN DETH, Jan W. and SCARBROUGH, Elinor (eds.): The Impact of Values. Oxford, Oxford University Press, 1995. SHAMIR, Michal: «Are Western Party Systems ‘Frozen’? A Comparative Dynamic Analysis», Comparative Political Studies 17/1984, p. 35-79. SHIVELY, W. Phillips: Review of Culture Shift in Advanced Industrial Society. Journal of Politics 53/1991, p. 235-238. SMITH, Gordon: «A System Perspective on Party System Change», Journal of Theoretical Politics 1/1989, p. 349-363. TORCAL, Mariano: «La dimensión materialista-postmaterialista en España: las variables del cambio cultural», Revista Española de Investigaciones Sociológicas 47/1989, p. 227-254. TORCAL, Mariano: «Análisis dimensional y estudio de valores: el cambio cultural en España», Revista Española de Investigaciones Sociológicas 58/1992, p. 97-122. VAN DETH, Jan W.: «A Macro Setting for Micro Politics» in VAN DETH, J. W. and SCARBROUGH, E. (eds.): The Impact of Values. Oxford, Oxford University Press, 1995. VAN DETH, Jan W. and SCARBROUGH, Elinor: «The Concept of Values» in VAN DETH, J. W. and SCARBROUGH, E. (eds.): The Impact of Values. Oxford, Oxford University Press, 1995. WEISBERG, Herbert F.: «Dimensionland: An Excursion into Spaces», American Journal of Political Science 18/1974, p. 743-776. (Also published in ASHER, H.B. WEISBERG, H.F. KESSEL, J.H. and SHIVELY, W.P.: Theory-Building and Data Analysis in the Social Sciences. Knoxville, The University of Tennessee Press, 1984).