Does Belgium (Still) Exist? Differences in Political Culture between Flemings and Walloons

West European Politics, Vol. 29, No. 5, 912 932, November 2006 Does Belgium (Still) Exist? Differences in Political Culture between Flemings and Walloons JAAK BILLIET, BART MADDENS and ANDRE -PAUL FROGNIER The emergence of two separate political systems in Belgium is the logical consequence of a gradual cultural and social divergence between Flanders and Wallonia. It is particularly in Flanders that Belgian nationhood is being put under pressure by a subnational identity. At the same time, there are still widespread emotional ties to Belgium, albeit stronger in Wallonia than in Flanders, that act as an important counterbalance to both the social and institutional centrifugal forces. Church involvement was traditionally higher in Flanders, but the stronger decline in the latter region has levelled off this difference. In both regions, the impact of pillarisation on voting behaviour is diminishing. As concerns the attitude towards foreigners, Flemings are more concerned about maintaining their own culture, while the Walloons are more concerned about the alleged economic and social threat. Other differences with regard to values exist, but are relatively small from a comparative perspective. Belgium s neighbours sometimes look at Belgium with some curiosity. Does that small country where Europe s capital is based really still exist? Many observers are under the impression that Belgium is in the process of falling apart in the wake of repeated constitutional reforms. Some have the impression that Belgians find it difficult to participate in any kind of activity as one nation. For instance, in its representation in the standing committees of the European Science Foundation, Belgium is the only country to send a delegation from both language communities. In the European Social Survey, Belgium is also the only country incapable of compiling its data under the leadership of one national coordinator. Is this image of a slow disintegration played out in reality? Or is Belgium a country capable of maintaining its unity despite its cultural differences, thanks to the high level of autonomy of its composite parts and the original solutions it has found? The topics discussed in this article touch upon this question. In a number of areas and developments, we go in search of what divides Belgians and Correspondence Address: jaak.billiet@soc.kuleuven.be ISSN 0140-2382 Print/1743-9655 Online ª 2006 Taylor & Francis DOI: 10.1080/01402380600968802

Differences in Political Culture between Flemings and Walloons 913 what unites them. In so doing, we look not only at behaviour and beliefs as areas in which the inhabitants of the different regions are different or alike; we also look at some structural developments which may or may not favour the preservation of a political entity. The development of two separate party systems, together with the almost complete separation of cultural life the media in particular are certainly trends which are gradually dividing the country. A difference can also be seen in the extent to which the citizens of the different regions identify with the Belgian nation or the sub-nation. However, we can also see similarities. It has long been taken for granted that Flanders is Catholic and Christian democrat, Wallonia unbelieving and socialist, and Brussels liberal and free-thinking. At both ideological and philosophical levels, this stereotypical picture is increasingly being eroded. At the religious level, the differences between the regions are shrinking and de-pillarisation is increasing, although pronounced links still exist between social organisations and the political parties. Another stereotypical picture that points towards separation is the idea that Flemings are more rightwing and xenophobic, a notion that is nurtured by the existence of a large extreme right-wing party in Flanders. Francophones, then, are supposed to be left-wing and much more open, but they allegedly have less faith in political institutions and the government. We will see, however, that the reality is much more subtle. This is the common thread running through this discussion. The Development of Two Separate Party Systems As described more in detail by De Winter, Swyngedouw and Dumont (this volume), the most striking feature of the Belgian party system is that since the end of the 1970s, there is no longer a single party system, but two distinct party systems, one Flemish and one francophone. On the Flemish side, we currently find Christian-democratic CD&V (Christian Democratisch & Vlaams, previously called CVP), the social-democratic SP.a (Socialistische Partij anders, previously SP), the liberal VLD (Vlaams Liberalen en Democraten, previously PVV), the Greens (Groen, previously AGALEV), the separatist and xenophobic Vlaams Belang (previously Vlaams Blok, and finally the Volksunie (VU, which split in 2000 into the independent N-VA, Nieuw Vlaams Alliantie, and the postnationalist Spirit). The first five party families can also be found, although at different strengths, on the francophone side: the Christian-democratic CdH (Centre de mocrate humaniste, previously PSC), the social-democratic PS (Parti Socialiste), the liberal MR (Mouvement Re formateur, previously PRL), the Greens (Ecolo), and the xenophobic Front National (FN). Apart from Brussels and the surrounding area, Flemish and francophone politics form two different worlds. The media report only to a limited extent on the politics of the other region. It is unusual for a politician from another region to be interviewed or to participate in a political debate. Most citizens

914 J. Billiet et al. are therefore scarcely involved, if at all, in the politics of the other region and their knowledge of it is extremely limited (Dewachter 1996: 136). The Development of Two Separate Societies The emergence of two separate party systems is not, of course, an isolated phenomenon, but reflects a broader social development in which the Flemish and francophone regions are drifting further apart. This increasing divergence has led to a situation where the sense of common political community required for the continued existence of unitary political parties has gradually disappeared. Many factors have contributed to this social duality in Belgium. Firstly, the language difference, which for a long time was the main reason behind Flemish Walloon conflicts. Belgium has been able to function perfectly well as a bilingual nation. However, this was mainly because the Flemish political, economic and cultural elite was largely francophone until the nineteenth century and, subsequently, increasingly bilingual. The Flemish elites remained closely involved in francophone social and cultural life, read the French-language newspapers and communicated fluently with their largely monolingual French-speaking compatriots. For the middle classes, too, knowledge and use of French were taken for granted. This gradually became less true as Flanders following the language legislation became officially monolingual Dutch-speaking. The youngest generations of Flemings, unlike their parents and grandparents, have grown up and become socialised in a monolingual Dutch environment. Moreover, they are much more focused on the Anglo-Saxon world, which means that French has literally become more of a foreign language. The most tangible consequence of this development is that virtually no French-language newspapers are read in Flanders and virtually nobody there watches francophone Belgian television channels. Likewise, very few Walloons or francophone Brussels residents follow the Flemish media (Dewachter 1996: 134 5). The three communities each have their own stations (VRT for the Flemish Community, RTBF for the French Community and BRF for the German Community), while the commercial radio and television stations are also monolingual. Primarily as a result of this media gap, two different cultures have gradually emerged, with diverging social sensitivities, fashions and customs. This trend towards cultural divergence was institutionalised and, at the same time, enhanced by the subsequent reforms of the state. Cultural policy was one of the first competences, as far back as 1970, to be devolved to the communities. In 1989, the communities were also given authority over education. This brought about a separation between Flemish, French and German education systems. Even in bilingual Brussels, the French-speaking and Dutch-speaking education networks are entirely separate. This means that the central Belgian authority has barely any policy instruments to

Differences in Political Culture between Flemings and Walloons 915 promote or socialise a shared Belgian culture, assuming the existence of a political will to do so. The fact that citizens know nothing about the other region and are becoming increasingly alienated from it, leads logically to a decline in interregional mobility. Relatively few Flemings or Walloons look for jobs in the other region, which prevents optimum harmonisation between demand and supply on the Belgian labour market and increases unemployment (Nationale Bank van Belgie 2004: 57 9). Related to this, the number of mixed Flemish Walloon marriages is also very low. In 1991, only 7.4 per cent of men and 7.6 per cent of women in Flanders had a marriage partner from another region. In Wallonia, these percentages are a little higher 19 per cent and 15.3 per cent respectively but this also includes those Walloons whose marriage partner comes from the largely francophone Brussels Region (Bartiaux and Wattelar 2002: 91 5). The Identity of Flemings and Walloons Due to the unconsolidated process of nation-building in Belgium, identification with the nation has always been relatively weak, even before the start of the federalisation process. As a result of the successive constitutional reforms, moreover, the Belgian nation has met with competition from the Walloon and Flemish entities, which also appeal to the citizens sense of identity. The Flemish government, in particular, occasionally pursues a genuine policy of nation-building, aimed at creating a Flemish national consciousness. In this way, the new Flemish institutions to some extent take over the role which was long played by the Vlaamse Beweging (Flemish Movement). This policy is also a way for the Flemish government to legitimise itself. On the French-speaking front, this desire for a (sub-) national sense of identity is much less obvious. This is because, on the one hand, Belgian identity is seen as less problematic than in Flanders, but also because no uniform object of identification is present. On the Flemish side, the Flemish Region and the Flemish Community have merged, as a result of which a single sub-nation has emerged, with one government and one parliament. By contrast, the Walloon Region and the French Community form two separate institutions. Advocates of a subnational consciousness are fighting different corners: some prefer to advocate a Walloon regional consciousness while others primarily want to promote identification with what is known as the Communaute Wallonie-Bruxelles. This division at the levels of both the institutions and the policy visions is one of the reasons why subnational or regional consciousness is relatively limited in Wallonia. The more recent survey results using the Moreno question indicate that a large proportion of Walloons (39.1 per cent) attaches equal importance to Walloon and Belgian identity (Table 1). For a relative minority, however, Belgian identity is to a greater or lesser extent more significant. Compared to this, the group for which Walloon identity is more important is very small

916 J. Billiet et al. (11.9 per cent). In Wallonia, Belgian feelings are somewhat stronger among young people and the more highly educated but, in general, the correlation between national consciousness and background characteristics is weak (De Winter 2003). If we look at the development over time of territorial identities (Table 2), using the question With which level do you identify most?, we find that in Wallonia the absolute majority that identifies with Belgium increases steadily, while identification with the region or community continues to decrease. In Flanders, Belgian national consciousness has been put under pressure most by a form of subnational consciousness, although we do notice an important shift over time. While before 1986 most Flemings identified in the first place with Flanders, Belgium has become the first choice since 1986, and for more than an absolute majority in the last decade. Here too, it does not seem to be true that the Belgian sense of identity is gradually being eroded as TABLE 1 IDENTIFICATION WITH BELGIUM AND THE REGION, IN FLANDERS AND WALLONIA, 2003 (VERTICAL PERCENTAGES) Flanders Wallonia Only Fleming/Walloon 7.4 3.6 More Fleming/Walloon than Belgian 23.9 8.3 As much Fleming/Walloon as Belgian 42.8 39.1 More Belgian than Fleming/Walloon 14.0 18.2 Only Belgian 11.9 30.8 N 1204 731 Source: ISPO/PIOP, 1999 (samples weighted by age, gender, education and vote). TABLE 2 FEELINGS OF BELONGING IN FLANDERS ON BASIS OF HIERARCHICAL QUESTION IN THE 1979 2003 PERIOD (PERCENTAGE OPTING FOR A SPECIFIC LEVEL AS FIRST CHOICE) 1979 1980 1981 1982 1986 1992 1996 1999 2003 Wallonia Belgium 54.3 59.4 61.3 57.4 57.6 69.0 71.4 73.3 67.9 Francophone Community 15.7 16.2 17.0 15.6 10.7 11.4 9.2 4.9 6.1 Walloon Region 18.4 17.4 15.0 17.7 15.7 12.2 10.0 10.7 10.3 Province 2.0 0.7 1.0 1.4 2.2 0.0 0.7 1.7 2.2 City/commune 9.4 6.3 5.8 7.8 11.9 7.4 8.7 9.2 10.1 Flanders Belgium 34.35 40.12 37.26 34.81 43.89 43.09 53.42 56.9 53.8 Flemish Community/ 44.99 44.44 47.36 48.52 32.57 40.53 25.59 27.3 29.3 Region Province 1.94 1.85 2.16 2.22 3.17 3.07 3.25 2.1 3.6 City/commune 18.71 13.58 13.22 14.44 20.36 13.31 17.71 13.5 12.3 Source: ISPO/PIOP, 1999 (samples weighted by age, gender, education and vote); Regioscopes I to VI (Delruelle and Frognier 1983).

Differences in Political Culture between Flemings and Walloons 917 the federalisation process continues. Also, if we look at the Moreno question (Table 1), most Flemings do not regard Belgian and Flemish identities as conflicting, but as complementary. In 2003, a relative majority of 43 per cent felt just as much Fleming as Belgian. In addition, however, another 31.3 per cent felt more or exclusively Flemish, compared to a somewhat smaller group of 25.9 per cent who felt more or exclusively Belgian (De Winter 2002). In other words, the group with a dominant regional or subnational identity is significantly larger than in Wallonia. In Flanders, a clear link exists between level of education and national consciousness, but it is the reverse of the situation in Wallonia: the higher the educational level, the more a person identifies with Flanders. By contrast, national consciousness is not linked to age (De Winter 2002; Maddens et al. 1994). In the Brussels Region, we find more or less the same pattern as in Wallonia: identification with Belgium prevails, albeit to a slightly lesser extent. The fact that Brussels is home to many European institutions and is often regarded as the capital of Europe is reflected in a stronger sense of European identity among the inhabitants of Brussels. Thirty per cent of these strongly identify with Europe as opposed to only 21.9 per cent in Wallonia and 12.8 per cent in Flanders (Billiet et al. 2000: 24). At the same time, these clear differences between Flemings and francophones should be put into perspective, taking into account that in Flanders, Wallonia and Brussels identification with the country or region occupies only a subordinate place in the broader sense of identity. By far the most Belgians (36.1 per cent) identify first and foremost with a primary group (such as family, friends or neighbours). This is followed by ideological groups and associations, with which 15.3 per cent and 9.8 per cent respectively feel most closely linked. Political entities, such as country, region or community, only rank fourth, with 5 per cent (Billiet et al. 2000: 18 19). The degree of national or subnational consciousness is, of course, closely entwined with the political attitude to constitutional reform and the level of autonomy for the provinces. This is why the stronger subnational feelings in Flanders are also translated into more outspoken support for the further federalisation of Belgium. By the end of the 1990s, 60.7 per cent of Flemings were in favour of increasing the autonomy of the regions to a certain degree, within the Belgian state (Table 3). In fact, 33.8 per cent were in favour of dividing up social security one of the last supporting pillars of the unitary Belgium compared to 34.8 per cent against (Meersseman et al. 2002). Nonetheless, only a small minority of Flemings, about 12 per cent, is in favour of separatism and complete Flemish independence. Another minority of 14.5 per cent advocates a complete return to a unitary Belgium to the detriment of the regions, while 12.7 per cent wants to return some powers to the centre. It should also be pointed out that only a few Flemings regard the Flemish/ Walloon issue as a political priority and take it into consideration in their voting behaviour. In the 1999 elections, only 2.2 per cent mentioned it in the

918 J. Billiet et al. TABLE 3 ATTITUDE WITH REGARD TO THE DESIRED STATE STRUCTURE IN 1999, BY REGION (VERTICAL PERCENTAGES) Flanders Wallonia Return to unitary Belgium 14.5 46.7 Federal state with more powers for the central 12.7 13.9 authority than now Federal state with more powers for the communities 28.1 14.6 and regions than now Federal state with as many powers as possible 32.6 19.3 for communities and regions Separatism 12.0 5.5 N 1760 1112 Source: ISPO/PIOP, 1999 (samples weighted by age, gender, education and vote). broad sense as one of the reasons for their choice of a party (Swyngedouw et al. 2001: 7). With this in mind, it may be surprising that a radical separatist party is currently the largest in Flanders. However, Vlaams Belang draws voters mainly on the basis of the issues of immigrants and security and less on the basis of the Flemish/Walloon issue. At the same time, since the disappearance of the VU, this issue has become somewhat more important among the Vlaams Belang electorate. In Wallonia, those favourably disposed towards Belgium are significantly more numerous than in Flanders. No fewer than 46.7 per cent are out-andout advocates of a return to a unitary Belgium, while an additional 13.9 per cent want to maintain federalism, but with a strengthening of the central authority (Table 3). By contrast, 33.9 per cent are in favour of a strengthening of the regions within the Belgian state and a mere 5.5 per cent are separatists (De Winter 2003). In Wallonia, only 12.4 per cent are in favour of dividing up social security. This should come as little surprise because it is particularly the less prosperous Walloon Region which has an interest in a federally organised social security system and also because the Walloon political elite is strongly against splitting it up. On the whole, it appears that, despite the growing social duality referred to above, a Belgian feeling still exists and has actually been on the rise in recent years. This should be seen as a latent undercurrent in public opinion, which can be sporadically activated by emotionally charged events. For example, whenever Belgium does well in the football World Cup or if a Belgian final is played at the Roland Garros tennis tournament, then the Belgian tricolour is flown proudly and everyone is Belgian again, rather than Flemish or Walloon. Following the Dutroux case in 1996 and the ensuing white march in Brussels, a strong feeling of national solidarity was evident. However, the most pronounced demonstration of Belgian feelings took place in 1993 following the death of King Baudouin. The fact that the Belgians mourned their deceased monarch en masse was regarded by

Differences in Political Culture between Flemings and Walloons 919 observers as a sign that Belgian national consciousness is still widespread and deep-rooted, whatever form it may take. The latter also provides a good illustration of the importance of the monarchy as a symbolic bond between Flemings and Walloons. Around 52 per cent of Belgians in 2003 agreed with the statement, we should be happy that we have a King because the country would fall apart otherwise. Here too, however, major differences are evident between Flanders, where only 43 per cent are of this opinion, and Wallonia, where 63 per cent subscribe to this view. As Belgium is institutionally dismantled and social duality increases, the role of the monarchy as a unifying factor actually seems to be growing. This is facilitated by the fact that the royal house can rely on considerable popularity throughout the country. Nonetheless, the monarchy is not entirely immune to the growing dualism in Belgian society. For a long time, it was particularly in the predominantly socialist Wallonia that the monarchy was contested, while the more conservative Flanders took up the defence of the monarchy. This contrast was most clearly expressed during the post-war royal question. In a referendum on the return of Leopold III to the throne in 1950, a majority of Flemings voted in favour and a majority of Walloons against. However, in recent decades, the roles seem to have been reversed and support for the monarchy is more widespread in Wallonia than in Flanders. In Flanders, we observe that support for the monarchy decreased during the 1990s: in 1990, 62.1 per cent of Flemings still believed that Belgium needed a King but by 2003 that had shrunk to 52.7 per cent. In Wallonia, the proportion responding positively to this question is currently significantly higher than in Flanders, at 71 per cent, while it was comparable in 1990 (59.6 per cent). This is perhaps related to the fact that the monarch, in strongly defending the unity of the country and explicitly rejecting the more radical demands for increased regional autonomy, fits in more closely with Walloon than with Flemish public opinion. This is also why Flemings sometimes have the impression that the monarch serves primarily francophone interests. For instance, in 2001, 46.3 per cent of Flemish 18 year olds agreed with the statement that the King was much closer to francophones than to Dutch-speakers, compared to only 17 per cent who rejected this statement (Maddens et al. 2002: 559). Secularisation in Belgium: Increasing Similarity? Belgium is known as a Catholic country. In 1846, the first census in Belgium registered 99.8 per cent of the population as Catholic. By 2003, according to the European Social Survey (ESS), this figure was only 43 per cent of the population aged over 15. In Flanders, half of the population still refers to itself as Catholic but in Wallonia and Brussels only one in three inhabitants still professes membership of the Catholic Church. Of course, this comparison between 1846 and the present does have its shortcomings. In 1846, the survey dealt with those actually baptised Catholic, while the ESS relates

920 J. Billiet et al. to subjectively considering oneself Catholic. For this reason, we would do better to compare the present figures with similar surveys. In 1981, the European Value Study (EVS) still registered 72 per cent Catholic. Less than ten years later, in the EVS for 1990, this was only 55 per cent, after which it declines to a minority of around 43 per cent in 2003. At the same time, half of the Belgians included in the ESS claim that they do not (or no longer) belong to any church. This places Belgium, of the 22 countries that participated in the ESS, in a central position between Greece and Poland as the most religious countries and the Czech Republic and Sweden at the other extreme. These data are in line with a recent study by the sociologist of religion, Dobbelaere; using the church statistics, he reached approximately the same conclusion regarding weekly church attendance (around 10 per cent in 1998). He also observes that little difference is evident among the regions (Dobbelaere 2003: 10). Until the 1980s, major differences still existed among the populations of Flanders, Wallonia and Brussels. Dobbelaere also demonstrates that the proportion of marriages still celebrated in church has fallen to 50 per cent. Christian burials have also declined, but the percentage is still around 80 per cent. This can partly be explained by differences in age as far as church involvement is concerned. In the generation aged over 60, according to the recent ESS data for Belgium, 31 per cent still go to church every week, but this figure drops to 9 per cent among those aged 45 60. In the age groups under 45, weekly religious practice is below 4 per cent. In this age group, the number of inhabitants of Belgium who never participate in religious services and who do not regard themselves as religious is well over 50 per cent. Among 45 60 year olds, 42 per cent do not belong to any religion but in the category aged over 80 this figure is only 27 per cent. This indicates that secularisation is taking place chiefly among the generations that have not yet reached the age of 60 (in other words, baby-boomers). If we assume that abandoning the church is mainly a phenomenon affecting young adults, this would fit in with the findings of Dobbelaere (2003), which place the most rapid decline in church involvement during the 1960s and 1970s. The decline took place more quickly in Flanders than in the rest of the country so that, as stated above, the differences between the regions in terms of church involvement have diminished. In the 2003 electoral survey, the figures for weekly religious practice are so similar that no significant difference can be seen: 9.5 per cent in Flanders, 8.7 per cent in Wallonia and 10.9 per cent in Brussels. The fact that Brussels is now at the same level is related to the presence of Muslims. Of course, weekly participation in religious services is not the whole story. We know from research that many people still take part in public religious services regularly, or at least during the major religious festivals. As far as Belgium is concerned, in addition to the more than 10 per cent of weekly practising believers, another 8 per cent participate regularly and 15 per cent only take part in services on major religious Holy Days (e.g. Christmas, Easter, All Saints). In addition, many more people come into contact with

Differences in Political Culture between Flemings and Walloons 921 religious services mainly within the context of funerals. It could be said that, at such times of loss, most people still come into contact with the church. Secularisation, De-pillarisation and Voting Behaviour Despite growing secularisation among Flemings and francophone Belgians, church involvement continues to exert a strong influence on voting behaviour for certain parties. Thanks to successive electoral surveys, we have a reasonably good picture of the composition of the electorates of political parties by church involvement and membership of social organisations in the various regions, spanning a period of 13 years. Table 4 shows the net effects of voting for each party, in the 2003 parliamentary elections, depending on the level of church involvement. The column percentages on the right (total) give the distribution of all votes cast by the voting populations of Flanders and Wallonia. The row at the top of each table section gives the distribution of these populations by church involvement. We present the effects as differences between the average percentage of votes for a party and the net percentage of votes for that party within each category of voters. The net percentages (or odds) of votes for a party are usually a good deal lower than the observed percentages because the other characteristics, such as age and health insurance fund, have been removed. In the table, all effects of more than five percentage points are given in bold. This makes it easy to see where a party has scored considerably below or above its average. TABLE 4 NET EFFECTS OF VOTING FOR EACH PARTY BY CHURCH INVOLVEMENT IN FLANDERS AND WALLONIA IN 2003, EXPRESSED AS DEVIATIONS FROM THE GRAND MEAN OF EACH PARTY Flanders 2003 None (11.1%) Free thinker (6.7%) Non-practising (46.6%) Irregular churchgoer (20.7%) Regular churchgoer (11.9%) Total* Agalev þ0.47 þ4.92 70.13 þ0.45 72.38 3.9% CD&V 712.18 716.85 74.09 þ10.12 þ25.82 20.9% N-VA þ0.44 72.95 70.60 þ1.09 70.54 4.6% SP.A-Spirit þ10.90 þ11.03 71.07 74.54 73.59 23.6% VLD þ1.73 73.72 þ4.07 74.86 711.22 24.3% Vlaams Blok 71.59 70.29 þ1.53 þ3.78 710.29 17.9% Wallonia 2003 None (13.4%) Free thinker (11.7%) Non-practising (42.0%) Irregular churchgoer (16.9%) Regular churchgoer (11.8%) Total* CDH 79.94 710.30 71.78 þ7.40 þ17.20 15.2% Ecolo þ5.85 71.88 73.59 þ1.99 71.05 7.3% FN 74.95 þ10.01 þ3.54 75.07 74.92 5.5% MR 72.53 713.47 70.30 þ6.62 þ8.62 28.0% PS þ5.27 þ9.27 þ4.40 75.82 718.22 36.1% ISPO data: N ¼ 1.213 in Flanders and 777 in Wallonia. *Calculated based on all voters, blank and spoiled votes included, but these are not shown in the table. Those of other religions are not included because of their low numbers.

922 J. Billiet et al. Involvement in church life in Flanders remains an important predictor of voting for the CD&V, although the overall influence of religious involvement is declining. Table 4 shows that, in Flanders, the net odds of voting for the CD&V in 2003 is indeed more than twice as high as the average (0.474 compared to 0.205). Furthermore, it is striking that being of no religion or liberal (free-thinker) has a strong effect on the chance of voting for SP. a-spirit or for Agalev. The VLD has the highest net odds among non-practising voters. The net odds of a vote for the Vlaams Blok is the smallest among devout voters. We know from the ISPO surveys 1 that, until 1999, the Vlaams Blok had more success among non-believers and, in 1991, among liberals (Billiet et al. 2000) but, in 2003, this seemed to have changed. For the first time, the Vlaams Blok probably obtained more votes than the average from the (irregularly) practising Catholics, while that chance is no higher than the average among voters who claim no religious conviction and among liberals. In Wallonia, the odds of voting Front National is apparently much larger among voters who refer to themselves as liberal. It is still true that, in Wallonia, being of no faith or liberal still strongly favours the socialist vote and prevents voting for the Christian democrats. Voters in Wallonia who regularly go to church are only half as likely to vote for the PS than the average odds in the electorate as a whole. The exact opposite is true for devout Catholics. In Wallonia, the MR has considerably higher odds among practising and devout Catholics. The effects on a vote for the CDH are even higher but the net odds of voting for the MR are higher among these Catholics than the net percentage of a vote for the CDH. These data indicate that de-pillarisation, in the sense of the separation of voters from their traditional, pillarised party, had advanced further in Wallonia in 2003 than in Flanders, if the criterion of the relationship between church involvement and (reported) voting behaviour is used. A similar trend towards de-pillarisation, with some differences in pace between Flanders and Wallonia, depending on whether we look at sociological Catholics or socialists, can also be seen in membership of a trade union and health insurance fund. The declining predictability of voting behaviour based on membership of socialist organisations is advancing more rapidly in Flanders than in Wallonia. This can be deduced from the differences in odds of members of a Christian or socialist trade union, or of Christian or socialist health insurance funds of voting for the Christian democrats or socialists respectively, depending on the generation to which they belong (under or over 45). 2 In both the younger and older generations, the link between membership of a trade union and the political party associated with it is stronger among socialists than Christian democrats. The ideological colour of a trade union perhaps remains a predictor of voting behaviour, but de-pillarisation is felt more strongly among both Flemings and francophones in the generations aged under 45 because, on average, fewer votes are cast for Christian democrats or socialists. Among the youngest generation of members of the Christian trade union in Wallonia,

Differences in Political Culture between Flemings and Walloons 923 in 1999 the PSC no longer gained an advantage because the odds of voting for the socialists was equally high. This indicates that in Wallonia the PSC and, subsequently, the CDH are losing trade union votes to the socialists, a clear sign of de-pillarisation (Billiet 2006). Membership of health insurance funds shows a similar picture but, in Wallonia, this remains a very strong predictor of voting for the socialists, among both older and younger generations. In Wallonia, the socialist pillar remains firm (Billiet 2006). The increasing detachment of voting behaviour from membership of social organisations was associated in Flanders, between 1999 and 2004, with the growth of the Vlaams Blok, a party at odds with the traditional divisions on the electoral landscape. The Attitude of Belgians towards Newcomers In the special Eurobarometer of 1997, Belgium was presented as the country with the most racists among all European countries investigated at the time. Why is this? Thanks to the ESS, we currently have more reliable data about a large number of aspects of attitudes towards immigrants and asylumseekers. The picture is more subtle. We have been following the development of attitudes towards foreigners almost constantly since 1989. What are the main findings? One aspect that was repeatedly examined between 1989 and 2003 is the feeling of being threatened by the presence of immigrants in terms of employment, social provisions and culture. As far as this aspect is concerned, few differences exist between Flemings and francophones and not much has changed over the years. Around 1998, the economic threat had diminished somewhat but, subsequently, these feelings became as strong as they were 15 years ago. Almost six out of ten Belgians have the feeling that their social security is being threatened by the presence of foreigners and this feeling is stronger in Wallonia than in Flanders and Brussels. For instance, in 2003, four out of ten Flemings were of the opinion that immigrants who were long-term unemployed should be forced to leave, but this was less the case than in francophone Belgium, where half of the adult population is of this opinion. Things are slightly different in terms of cultural threat. Almost four out of ten Belgians feel threatened in their cultural individuality but that is just as much the case in Flanders as in Wallonia, but much less pronounced in Brussels. Where political rights are involved, such as the municipal voting right granted last year to foreigners who have been in Belgium legally for more than five years, various studies show that the vast majority of Flemings opposes this (between 60 and 65 per cent) while in Wallonia and Brussels around 45 per cent are against. More francophones (67 per cent) than Flemings (50 per cent) are of the opinion that people who come to live in Belgium should have the same rights as everyone else. Over 57 per cent of francophones claim to have some friends from another country; among Flemings this is considerably lower (37 per cent).

924 J. Billiet et al. In other words, quite a few differences are still evident between the regions depending on the specific aspect that is related to immigration. Flemings usually feel somewhat more threatened in their cultural individuality and are less likely to seek out social contact with foreigners. In francophone Belgium, particularly in Wallonia, where the economic situation is the most problematic, people feel most threatened at economic level and in terms of social provisions. The attitude towards asylum-seekers is another aspect that has been examined recently. Do Flemings have a different attitude towards asylumseekers than their southern compatriots? In order to answer this question, we use the data from the ESS. This face-to-face survey, conducted in 2002 among 1,899 Belgians, contains a large number of items about admission policy and about the subjects of immigration and asylum. Members of the French-speaking community seem to hold a more tolerant attitude than their northern neighbours. As far as admission policy is concerned, Belgians are, in general, not very favourably disposed towards admitting many people from outside Belgium, no matter what their characteristics the same or a different ethnic group, rich or poor. When forced to choose, people were more inclined to accept immigrants from poorer countries rather than people from richer countries. A minority of around 15 per cent simply does not want to admit anyone else. These findings provide a subtle picture of the admission or reception of foreigners. Most Belgians appear to be receptive to arguments in favour of accepting asylum-seekers, as long as their number is not too high. Most Belgians agree on a number of aspects of the asylum issue. For instance, 68 per cent of Belgians believe that Belgium has proportionately far too many asylum-seekers. The majority of Belgians (61 per cent) is also of the opinion that asylum-seekers should be granted permission to work in Belgium while their asylum applications are being processed. On this point, virtually no differences exist between Flemings and francophones. A narrow majority of 56 per cent does not agree that the government should be flexible when assessing asylum applications and does not advocate the right to family reunification for recognised refugees; this percentage is considerably higher in Flanders (over 67 per cent). Francophones are also somewhat more inclined to provide financial support (30 per cent compared to 19 per cent of Flemings) and to allow family reunification (36 per cent compared to 23 per cent) and only half as many francophones (20 per cent) as Flemings (40 per cent) agree with housing asylum-seekers in closed centres (Meireman et al. 2004). With respect to various aspects of asylum policy and immigration, the Belgians are not so different from the Dutch. They rank approximately in the middle of the 22 European countries that participated in the first round of the European Social Survey. The Nordic countries are generally more open and the southern and central European countries more closed towards immigration.

Differences in Political Culture between Flemings and Walloons 925 Social and Political Trust A final set of political views attracting our interest relates to confidence in the future and in others, including politicians and institutions. We prefer to report the figures of the 1999 election survey since the list of institutions was then much larger than in 2003. Significant changes in 2003 are reported in the text. Following the Dutroux case in the mid-1990s, confidence in institutions and in the police took a serious blow. Among all institutions, education and the King could rely most on the confidence of the Flemish population. In research following the 1995 elections, in other words before the infamous summer of 1996, over 70 per cent of Flemings still had a high or very high degree of confidence in education and 50 per cent in the King. In 1999 in Flanders, education still retained the same level of support, but confidence in the King fell considerably to 41 per cent (Table 5). In Wallonia, confidence in the King remained at more or less the same level and in Brussels it actually rose by five percentage points compared to 1995. Trust in the King rose more than ten percentage points by 2003 in Brussels and Wallonia. In Wallonia and Brussels, by the end of the 1990s, citizens had less confidence in education (around 55 per cent) than in Flanders and this remained reasonably stable between 1995 and 1999. This cannot be separated from the limited resources on which the French Community can rely as a result of the economic situation and the successive strikes suffered by the education system over this period. TABLE 5 CONFIDENCE IN SOCIAL INSTITUTIONS IN THE BELGIAN REGIONS IN 1999 (PERCENTAGES OF HIGH AND REASONABLY HIGH CONFIDENCE) Institution Flanders Wallonia Brussels Total Education system 72.6 55.0 52.2 65.5 Preferred political party ns 47.5 43.8 51.3 46.5 King 41.3 52.2 49.6 45.5 Federal police (Rijkswacht) 38.1 29.7 24.6 34.4 Banks 37.7 24.0 28.1 23.4 Federal Parliament 19.1 25.8 27.8 21.9 European Parliament 18.2 25.0 30.0 21.2 The press 16.3 28.1 23.6 20.6 Trade unions 18.3 20.3 22.5 19.2 The government 17.1 24.2 22.3 19.8 Church ns 18.3 20.3 16.7 18.8 Central administration 12.3 23.8 16.0 16.3 The judicial authorities 13.9 19.9 17.5 16.1 Employers organisations 11.7 14.3 12.9 12.7 Political parties ns 8.2 11.1 11.2 9.4 N 2,179 1,483 577 4,239 Note: ISPO/PIOP data from 1999 electoral research. All differences are statistically significant. ns Not statistically significant at level 0.01.

926 J. Billiet et al. Confidence in the justice system (judicial authorities) was already low in 1995 in Belgium (19 per cent) and fell even further in 1999 (16 per cent) and it remains low in 2003. The decline in confidence in the judicial authorities took place mainly in Flanders and Brussels. Confidence in the federal police (then still the Rijkswacht) also fell from 41 per cent in 1995 to only 34.4 per cent in 1999. The most striking aspect is that this decline in confidence in the federal police of over ten percentage points was evident primarily in Wallonia and Brussels. This is not surprising, since it was chiefly the Rijkswacht in Brussels and Wallonia that was involved in the investigative work relating to the missing children. In francophone public opinion, the police fell short in many respects. The general figure for 2003 is about the same, and there are no longer differences between the regions in trust in the federal police. It is noticeable that confidence in a number of institutions such as the federal parliament, the European Parliament, the press, the government and the central administration is higher in francophone Belgium than in Flanders. The fact that the European Parliament scores the highest in Brussels is, of course, related to the fact that much European activity takes place in the capital, which produces tangible advantages for many citizens. Confidence in the banks is higher in Flanders. Public opinion there is obviously less left-leaning. In the ISPO/PIOP survey 3 from 1999, confidence in the political parties scores poorly in general throughout the country but we see a sharp contrast with confidence in the preferred political party. Evidently, people still have confidence in the parties for which they vote. The trust in the preferred political party is somewhat higher in 2003, but confidence in the government or the federal parliament has remained almost unchanged. In addition to information about confidence in institutions, we have data about the images citizens have of politicians and the political system. In particular, Wallonia stands out for its very negative picture of political life. In answer to the question about spontaneous reaction to the word politics, in 1999 57 per cent of Walloon voters responded with distrust, aversion and boredom. In Flanders and Brussels, the figure is a good ten percentage points lower. Proportionately fewer Walloon than Flemish voters are satisfied with democracy. This remained fairly stable in 1996 and 1999. The electoral survey of 1996 showed that Walloon voters had more markedly negative feelings towards politicians and elections. More than Flemings, they felt that promises would not be fulfilled and that little consideration was given to the wishes of the voters. This dismal view of politicians improved a little in 1999 and 2003. Differences still exist between Flemings and francophones, but they are somewhat less pronounced than in 1996. The feeling of political impotence and distrust has always been more strongly held in Wallonia and Brussels than in Flanders, but this is not only the case for lack of confidence in politics. The allegation that most people

Differences in Political Culture between Flemings and Walloons 927 disappoint once you get to know them was endorsed by a quarter of Flemish voters in 1996, but in Wallonia the figure was considerably higher, at 38 per cent. The general feeling of distrust also seems to have grown in Flanders. In 1989, 42 per cent of adult Flemings agreed with the statement, Nowadays you really do not know whom you can trust. In 1996, this had risen to 59 per cent, but in Wallonia, in the same period, this figure was over 68 per cent. Deeper research shows that this view is closely correlated with an attitude that expressed confusion, the feeling of complexity and uncertainty about the future. This general feeling of social confusion was a little stronger in Wallonia around 1996 but, subsequently, it diminished again and francophone Belgium presented a less unfavourable picture. In 1999 and 2003, half of Flemings still expressed this general distrust of others but, in Wallonia, this had fallen to approximately 44 per cent. Apparently not everything evolves in the same direction. This observation is a good starting point for thinking about how we should understand these tendencies in differing public opinions. The Values of Flemings and Francophones The European Social Survey examines the ten universal values from Schwartz (1992, 2003), using 21 verbal portraits ; these are short descriptions in which the respondent has to indicate how strongly he or she identifies with each portrait. An attempt was made to measure more concisely the extensive instrument used to measure every value based on four items. These basic values are independence, universalism, charity, tradition, conformity, security, power, success, hedonism and meeting challenges. In Schwartz s research, these ten values can be traced back to two main dimensions: openness to change versus conservatism, on the one hand, and self-enhancement versus self-transcendence, on the other. Since the value set was severely limited, it is not possible to reconstruct the ten original values using measuring models in a way that is equivalent for all 22 countries in the ESS. During a presentation at the first conference of the European Association for Survey Research, researchers presented a model with seven values that is equivalent for all countries (Davidov and Schmidt 2005). This model is obtained by merging universalism with charity, tradition with conformity and power with success. We will use this model. Below is a brief description of each of the seven values. We then examine to what extent Flemings and francophones differ in average scores and whether the effect of the language group is still evident after controlling for other relevant characteristics.. Universalism and charity: believing it important that everyone in the world be treated equally and have equal opportunities, that people try to understand one another, that they care for the environment and that people devote themselves to the welfare of people around them;

928 J. Billiet et al.. Tradition and conformity: always following the rules, upholding the customs acquired during one s upbringing and always acting modestly and properly;. Power and success: attaching much importance to money and material success, being admired by others and acquiring influence and recognition so that others do what he or she wants;. Security: believing that it is important to live in a secure environment and knowing that one is protected by a strong state;. Hedonism: taking every opportunity to pursue pleasure and to enjoy and indulge oneself;. Independence: being creative and wanting to do everything in an original way, feeling free and making decisions independently;. Taking on challenges: always going in search of new things, looking for adventure as well as being prepared to take risks. These are descriptions to which people attach great importance and the exercise involves indicating, on a six-point scale, how strongly the respondent resembles them. All value scales consist of two to five indicators and are constructed so that they vary between 0 (least similar) and 10 (most similar). The internal consistency or reliability of these scales is not strong, but moderate to weak. This is because each description contains more than one idea and is fairly complex. 4 Table 6 presents a summary of the average scores for Flemings and francophones. The differences are statistically significant in six of the seven cases (0.01 at a level). In order to assess the size of the differences, we look once again at neighbouring countries. With the exception of traditionalism/conformity and the security value, francophones achieve consistently higher scores than Flemings. This is particularly true of the values of independence and being willing to take on challenges and adventures. Flemings score significantly lower on these aspects. Francophones are also on average somewhat more focused on enjoying life and they also attach greater importance to having material success and influence. As far as the values of universalism and charity are TABLE 6 AVERAGE SCORES (OUT OF 10) FOR SEVEN VALUES AMONG FLEMINGS (N ¼ 1245) AND FRANCOPHONES (N ¼ 651) IN THE EUROPEAN VALUES SURVEY OF 2002 Group Univ. & charity Trad., conform Power, success Security Hedonism Independence Challenges Francophones 9.97 6.46 5.22 7.22 7.22 7.65 5.74 Flemings 9.76 6.45 4.87 6.99 6.75 7.07 5.03 Difference 0.20 0.01 0.35 0.24 0.47 0.58 0.71 T value 3.02 0.08 3.83 2.46 4.78 6.77 6.39 prob. 4jtj 0.003 0.933 50.001 0.014 50.001 50.001 50.001