Resumé

In one way, Norway's local elections would seem to be a deviant case: a protest election with a landslide victory to a recently established neo-liberalistic party, greater volatility among the voters, and an unusually low turnout. However, the election survey and other electoral statistics reveal a continuation and reinforcement of social and regional trends in voting patterns, trends which have emerged during recent elections.

Passive and Active Protest

The local elections held on 14-15 September 1987 have been characterized
as a protest election. In order to scrutinize this statement, we need to
differentiate between a passive and an active protest.

Passive protest is reflected in a turnout unusually low in a Norwegian context: nearly every third voter was an abstainer. Evaluation of this figure must, however, take into consideration that the turnout in local elections is always lower than in general elections. During the post-war period (1945-1987), election turnout has been approximately 8.5 percent below that of general elections. All the same, in 1987 participation was extraordinarily In fact we have to go back to the local elections of 1945, some months after World War 11, to find a lower turnout.

If we consider the gender aspect, we have to go back as far as 1922 to reach a similar low level of participation among men (see Fig. I).¹ During the post-war period the difference in turnout between men and women has almost disappeared (in 1987: 69.6 percent for men and 69.3 for women). But one more fact has to be added. The age threshold for voting in 1922 was 23 years, while in 1987 it was 18 years. Turnout is usually especially low among the younger cohorts.

Among women, the decline in turnout is far from dramatic. In fact the
1987 turnout was a little higher than in recent elections.

Active protest was manifested in a landslide victory for a neo-liberalistic
party, the Party of Progress. The forerunner to this populistic party was
launched back in 1973 when proponents of the welfare backlash has their

heyday. Originally entitled 'Anders Lange's Party', it was established with inspiration from Mr. Glistrup's success in the opinion polls in Denmark. In fact its Norwegian leader, the late Anders Lange, was represented in the Storting two months before Mr. Glistrup's entered the Folketing.

Norway's 1987 local elections marked a reversal of certain trends and a continuation of others. Observers of electoral behaviour have argued that Norway is moving towards a two-party system. As a consequence of the forward surge of the Conservative Party from the mid-19705, support behind the two major parties has increased. In the last Norwegian general election (1985), the electoral strength behind Labour and the Conservatives was 71 percent - as against 73 percent in the general election of 1987 in the homeland of the two-party system, Great Britain.

Together, Labour and Conservatives amassed approximately only 60 percent of the vote in the 1987 local elections. But it ought to be added that in local elections the main parties tend to emerge with somewhat less support than in general elections. The point is, however, that the support was markedly lower than in previous local elections.

The result of the 1987 local elections meant a defeat for the two leading parties, which more than any others represent the Establishment. However, this result cannot be attributed to a left-wing anti-Establishment trend. For the parties to the left of Labour the result of the election was more or less the status quo. Since the electoral base of Labour was diminished, one might easily have expected electoral gains for the parties to the left of the

social democrats, but this was not the case. This underlines the character
of the 1987 election as a protest with a swing to the right.

The result of a recently established Green Party was indeed not promising. The Green Party, which is not yet a nation-wide party, ran for election in only two counties (Oslo and neighbouring Akershus) and a rather meagre number of votes was collected, 0.6 percent, with only one member elected to the county council. In this respect, Norway clearly deviates from other Scandinavian countries.²

Supporters of Green parties generally deny the relevance of the traditional line between socialist and bourgeois parties. Despite this declared position above the radical/conservative camp, most 'green' supporters usually belong to the left wing. A recent study of the first and largest Green party - die Griine in the Federal Republic of Germany - concludes with a characterization of the party as part of 'the New Left' (Burklin 1987, 123).

What we stress here is that the weakness of the Green Party in Norway
provides further evidence of the absence of a protest with a leftwards swing.

Election Results

Table 1 shows the share of votes for the various parties. In local elections there are two tickets (except in the capital city, Oslo): one for the municipal council and one for the county council. Before 1975, county councils were appointed by members of the local councils in the various municipalities in each county. Since 1975, county assemblies have been directly elected by the voters in elections where the constituency is the entire county, just as in general elections. In the traditional local elections the constituency is the municipality, rather than the whole county. Despite such differences the outcome of the two elections is quite similar, with only minor deviations.

Local and non-partisan lists classified as 'others' received more votes in the traditional local elections than in county council elections. Such local and non-partisan lists have been in steady decline during recent decades (Hjellum 1966) although a slight increase can be observed from 1983 to 1987.

The Party of Progress represents another example of a difference in electoral support between the two forms of local elections. The underlying reason for this is that as a relatively recently established party, the Party of Progress did not take part in all the 448 municipalities in the contest for seats on the municipal councils.

Let us turn to the electoral statistics for the county council election,
which is more comparable with general elections than the traditional local
elections. In Table 1 the 1987 local elections are compared with the 1983

local elections and the 1985 general election. Parties are divided into three
blocs: socialist, middle, conservative.

The decline of the socialist bloc is unambiguous and is primarily due to the loss of votes cast for Labour. A decline can also be registered for the middle parties in relation to the 1983 local elections. Compared with the 1985 general election, however, the result is the status quo. On the other hand, the middle parties usually do better in local elections than in general elections. In the long run, the decline of the middle parties is the steadiest trend to emerge from the electoral statistics. From 1945 to 1975 the middle parties gained around 30 percent of the vote, whereas in 1987 the figure was under 20 percent.

The outright winner is the conservative bloc. Ever since the mid-1970s the conservative bloc has profited from the decline of both the socialist bloc and the middle parties. In 1987 the Conservative Party suffered a serious setback, but the gains in votes for the Party of Progress were larger than the loss for the Conservative Party alone. The Party of Progress is no longer a minor party in relation to the Conservative Party: the ratio is now one to two.

However, there are also signs of weakness. The relation between the
Convervative Party and the Party of Progress is a hostile one, and internal
controversies make the conservative bloc vulnerable to attack.

We shall now focus first on the election campaign and then analyse

individual changes from 1985 to 1987. Finally, we will look at the social and regional trends in voting patterns partly with the help of an election survey conveyed by the Market and Media Institute (MMI) just after the election.³

The Election Campaign

Although the 1987 elections were local the campaign was to a certain degree coloured by events on the national scene. Some months previously, there had been a serious collapse in negotiations among the three leading bourgeois (Conservative Party, Christian People's Party and Centre Party) in their efforts to re-establish a governmental alternative to the current Labour government. This unstable parliamentary situation had its background in the results of the 1985 general election, which had had no clear loser or winner. The bourgeois coalition government with Mr. Kåre Willoch as Prime Minister was re-elected. But the government lost its legislative majority (77 to 78) and became dependent upon support from two MPs from the Party of Progress. As already indicated, the relationship between this party and the other bourgeois parties is not one of mutual trust. From the viewpoint of the Party of Progress the Willoch government was at the time merely the lesser of two evils. In April 1986 the Party of Progress supported Labour and overthrew the coalition government, thus paving the way for a Labour government.⁴

Since the majority in the Storting belonged to what traditionally has been called the bourgeois bloc, it was commonly expected that this Labour government would be a temporary one. The situation led to impatience in the bourgeois camp and attempts to re-establish a coalition government. The Centre Party (the former Agrarian Party) hesitated. But after having been promised increased financial support for the farmers, the Centre joined the Conservatives (the primary force behind the negotiations), and the Christian People's Party in order to establish a new government. But the Party of Progress refused to support this alternative, since it meant financial transfers to the farmers, a policy which the party strongly opposed.

The efforts to re-establish a coalition government did not meet with widespread support among voters from the bourgeois bloc. The events gave the impression of a search for power for its own sake. One effect of the unfulfilled promise of a new government was a decline in trust in politicians, as registered by opinion polls. Another effect was the recruitment of new supporters to the Party of Progress. Even though the Party of Progress had blocked the possibility of a new bourgeois government, the party soared upwards in the opinion pools.⁵

The result of any election will be influenced by the issues in the forefront

of the campaign. Social welfare issues topped the agenda. Characteristic of such issues is that nearly everyone agrees on the aims. Indeed it is difficult not to approve the demands for better care for the elderly, and 'hospital queues' are generally regarded as undesireable. Such welfare issues are cross-political questions. However, several opinion polls have documented that the voters have more confidence in the capacity of a Labour government than in that of a bourgeois government to solve such problems (Valen & Aardal 1983; Hellevik & Bjørklund 1985).

As regards financial issues, there were complaints from all parties over the high rate of interest, especially in connection with house-building mortgages. A unanimous demand was heard: lower the rate of interest. No one however, could, explain just how this aim might be reached. Labour was clearly in the least comfortable situation: as the opposition, the party had attacked the bourgeois government for being insensitive to the increased rate of interest. Once Labour was in governmental position, however, the promise of a lower rate of interest seemed to evaporate. In fact the rate of interest continued to increase even after Labour had come to power.

The leadership of the Conservative Party tried to make the tax question a major campaign issue. Tax cuts were the message. But this message provoked internal struggle in the party which eliminated the issue as a mobilizing factor. Instead a split was revealed in the party, and this harmed the unity of strength.

Politicians from the leading parties hesitated to make the question of refugees and asylum-seekers an issue in the election campaign. They argued that making this a question of party politics could further harm the situation of the refugees. Obviously, tactics also lay behind this strategy. Grassroots reports indicated hostility towards refugees. In such a situation, political statements from the top of the system defending the refugees might provoke party loyalties and result in diminishing support.

The Party of Progress, however, did make the refugee question into a political issue. It came to the media forefront after the leader of the party Mr Carl I. Hagen, cited a letter - which later was shown to be a fake - saying that Norway would be on the way to becoming a Muslim state unless the borders were closed.

The proportion of foreigners in Norway is in fact modest: 2.5 percent of the population, or some 120,000 persons. Most of these non-Norwegians are from the Nordic countries, from Northern Europe and the USA, with a way of life not so different from that of the Norwegian. Only a quarter come from Asia and Africa.⁶

The immigrant question was raised largely as a consequence of the
exponential growth in asylum-seekers. According to Norwegian law, people
from other countries can be granted entry to Norway if they document that

they are bona fide refugees and in danger of persecution if sent back to
their homeland.

In recent years, refugees from the Third World have been coming farther
north in Europe partly as a result of more restrictive new legislation in
other European countries.7

In Norway, the number of asylum-seekers exploded in 1987: 8,613 as against only 2,800 in 1986 and, before then, below the 1000 mark. The Norwegian reception system for refugees was accustomed to numbers quite different from the reality experienced in 1987. This system has to be reconstructed, and ultimately it becomes a question of more money, additional grants from the government. The refugee issue was placed in an economic context. For the Party of Progress, the main argument was that increased grants to refugees would mean worse conditions for 'our own' inhabitants, those of Norwegian birth. More precisely they supported 'our own' with real problems, the elderly and sick people in 'hospital queues'. The spokesmen for the Party of Progress stressed welfare issues in this way as opposed to demands from asylum-seekers. They also argued that in reality only a minority of the asylum-seekers were real refugees.

How the election campaign actually influenced the electoral result is difficult to say. In one way it is dangerously easy to explain the election result ex post facto. But there is no doubt that the collapse in bourgeois governmental negotiations gave the Party of Progress a good starting point for its election campaign. The party went on to play a successful role in the campaign, with one of the mobilizing issues clearly being the immigrant question.

The cross-political character of both welfare issues and complaints about the high rate of interest made it difficult to identify between friends and enemies. The two major parties (Labour and the Conservatives) which have played a leading role in governments in recent years felt some responsibility for the inadequacies of the public health service and for the difficulties in the economic sphere. They had to face the bad effects of incumbency. This would seem a reasonable explanation for the failure of the two parties in the 1987 election.

Table 2 shows empirical support for our suggestions. In Norway, the municipality has responsibility for a range of duties. Respondents were asked if they felt that specially mentioned duties should have higher priority.. Answers to the question about welfare issues revealed the cross-political character: massive support across party lines.

In an opinion poll conducted by MMI, a question relating to the decisive issues for party preference — local or national issues — revealed that voters for the Party of Progress were in a special situation. This was the only party where a majority said that national issues were the most important ones. Behind this constellation we can perhaps see the influence of events on the

national scene: the collapse in the governmental negotiations and the
refugee question.

There was also a question about house-building, with only an indirect connection to the high rate of interest in financing house-building. But house prices are partly dependent on the balance between supply and demand. If the municipality took the initiative to build more houses, prices might perhaps fall. There is, however, no clear pattern between the answers as to this issue and party preference, an indication of its cross-political character. This is the case when the question is raised about the immigrants. Approximately a quarter of the voters want higher priority to problems related to immigrants. But this proportion is substantially lower among supporters of the Party of Progress. Their special position is confirmed when we ask about attitudes to immigrants. Foreigners from various countries were mentioned, but in Table 2 we have shown only attitudes to Pakistanis. In Norway, the Pakistani is seen as the most typical foreign worker from the Third World. Attitudes are definitely not without hostile elements, especially among supporters of the Party of Progress.

Individual Changes 1985 to 1987

The winner of the election - the Party of Progress - recruited supporters
from voters with and without previous party affiliation.

Mock elections at a large number of secondary schools indicated that this party had strong support among first-time voters, especially among the young men. In these mock elections, 25 percent of young men against 15 percent of young women voted for the Party of Progress. The total share was 20 percent compared with approximately 25 percent for Labour and the Conservatives.⁸ Not only the mock elections but also the election survey confirmed that among young men (under 30 years) the Party of Progress was the second largest party after the Conservatives.

The Party of Progress also recruited voters from among former supporters of other parties. A panel study from the Central Bureau of Statistics for the Norwegian Electoral Programme enables us to identify different votershifts the 1985 general election to the 1987 local elections.⁹

Stable voters from 1985 to 1987 fall in the marked cells along the diagonal in Table 3 and comprise 64 percent of the sample.¹⁰ Thus, 36 percent have shifted party or have moved in or out of the group of abstainers. This interelection between general and local elections was somewhat higher than had been observed earlier. Between 1973 and 1975 the proportion of unstable voters was 33 percent (Valen 1976). In recent elections the proportion of 'shifters' between parliamentary elections has been around one third of the total electorate (Valen & Aardal 1983). The volatility from 1985 to 1987 exceeded this proportion.

The balance between voting and non-voting influenced the election result. Only the Party of Progress seems to have recruited more votes from the group of previous stay-at-homes than its losses among the 1983 supporters who abstained from voting in 1987. The decline in turnout harmed all the other parties - but to varying degrees. Labour and the Conservatives suffered severe losses due to this group of abstainers. In the panel, 78 percent of Labour's losses from 1985 to 1987 were due to non-voting among former supporters. The corresponding figure for the Conservatives was 43 percent.

Table 3 reveals a shift of votes among all the parties. The important question is whether the exchange was positive. Did the party receive more votes than it lost? For the Party of Progress the exchanges were always balanced in a positive way; for the Conservative Party they were always negative. In Table 4 we have decompounded the gains to the Party of Progress and the losses to the Conservatives.

A stream of votes went from the Conservatives to the Party of Progress: 59 percent of the gains and 35 percent of the losses are due to this stream. The second largest contribution to the Party of Progress (24 percent of the gains) came from Labour. This result confirms a familiar pattern. The Party of Progress has always recruited most of its votes from the Conservative Party; Labour is ranked as number two.

Social and Regional Trends

Cleavages and Generations: Some cleavages are of growing importance (alignment) while others are of diminishing importance (de-alignment). This partly depends upon which cohort or generation is the unit of analysis, as different generations mean different cleavages.¹¹

Election surveys reveal that in recent decades some cleavages have gradually become more clear-cut than before. This is the case for the gender gap (Listhaug, Miller & Valen 1985; Bjørklund 1986), and for the working affiliation in the private versus public sector. Other cleavages, such as classvoting, been less obvious. The question is: do the local elections of 1987 confirm this long-term picture, or not?

The 'generation' concept can be useful for analysing the weakening or reinforcement of social trends. A political generation is based on common experiences in an early phase in life where impressions can give a stamp of lasting character (Mannheim 1972). In order to divide the electorate into distinct political generations we have to identify historically important events. In Norway, World War II is such a watershed. We may draw a dividing line between pre-war and post-war generations. Half of the current electorate (in fact 51 percent) was born after World War II; this consists in our terms of two post-war generations. The first post-war generation was born from 1946 to 1960. The core group here is the immediate post-war baby boom, which filled up the universities around 1968 and formed the 'student revolution'. The second post-war generation (born 1961-1969), by contrast, entered the electorate at a time when the conservative surge was just starting.

The first post-war generation in Norway has grown up in a society where social progress has been taken for granted. The second post-war generation has perhaps tended to take the affluent society for granted, but they have also met with problems; for example, the escalating prices on the housing market.

Between the first post-war generation and the pre-war generation we find the inter-war generation, born from 1926 to 1945. Representatives of this generation were children/teenagers during the wartime occupation of Norway. The pre-war generation (born before 1925) on the other hand had the responsibility for first bringing the nation through the war and then for the rebuilding phase after 1945 - the 'golden age' of the social democratic era. Indeed, it is in the pre-war generation that Labour still has its strongest and most loyal voter segment.

The Gender Gap in Different Generation: Recent years have seen a gender gap in voting behaviour. Young women swing to the left, young men to the right. Today, young men represent a stronghold for the conservative surge.

Among the voters in the 1987 election, the socialist bloc had a narrow majority of women (+2 percent unit) and the conservative bloc a clearer majority of men (+8 percent unit). The middle parties had more women than men among their voters mainly as a consequence of the traditionally strong appeal to women of the Christian People's Party.

In recent decades, the middle parties have been fairly stable in terms of the ratio between men and women, but this is not the case for the socialist and conservative bloc. In the elections just after World War 11, the conservative attracted more women than men, whereas on the socialist wing an opposite pattern could be observed (Duverger 1955). The current situation has in fact reversed these trends.

In Fig. 2 the electoral strength behind the socialist bloc (Labour and Socialist Left/Communist) and the conservative bloc (Conservatives and Progress) is shown separately for men and women.¹² The sample from MMFs election survey is divided into four generations (second post-war, '1968', inter-war and pre-war generation). In the pre-war generation the same pattern emerges as was observed in the surveys just after the war: women are more conservative than men. This is definitely not the case in the youngest generation, however.

Support for the conservative bloc among the youngest women does not differ much from what can be observed among women in the three older generations. Such stability across generation borders is not the case for men, however the electoral strength behind the conservative bloc is more than twice as strong in the youngest generation (second post-war generation) it is in the oldest (pre-war generation). The only case where the conservative bloc is stronger than the socialist bloc is among the youngest men (61 percent against 28 percent).

Young men form the most solid foundation for the Conservatives, whereas old men (pre-war generation), together with women from the '1968' generation, are at the opposite end as the most radical group: every second voter here supports the socialist bloc.

Figure 2 also gives information about the generation profile of the various
parties. Labour has its best support in the inter-war and pre-war generation.

The conservative bloc experienced internal changes in the 1987 local elections. As a result of the landslide victory of the Party of Progress, Conservative support among the post-war generations was weakened. The generation profile of the Conservative Party has subsequently become more diffuse. The Party of Progress on the other hand shows a very clear profile. This is a party for voters born after the war, the core group being men from the youngest generation.

The Party of Progress is not the only party to be launched in the postwarperiod.
new party is the Socialist Left, which also has a postwarprofile.
there is a difference. Strongest support for the Socialist

Left is in the '1968' generation, and not the youngest generation. In
addition, the Socialist Left today holds more attraction for women than for
men - the reverse of the situation of the Party of Progress.

Public/Private Sector: During the 1960s and 1970s the public sector expanded. Currently about one third of the labour force is employed in the public sector. The public sector is of growing importance in another sense, too: the dividing line between public and private sectors has been clearer in the voting patterns.

The old adage has it that blue-collar workers vote for socialist parties, and white-collar workers for bourgeois parties. This has to be revised. Among voters from the post-war generations the distinction between blue and white collar, or working class versus middle class, is overshadowed by the sector dimension (public/private). If we wish to predict the party affiliation (socialist/conservative) among wage earners in the post-war generations, then information about sector - public or private - is more important than the blue- versus white-collar dimension.

In the generations born before 1945 the old pattern still holds (see Fig. 3).¹³ Concerning support for the socialist bloc, the difference between blue and white collar is significant, while sector affiliation has no effect at all. Looking at the conservative bloc, however, we see that sector affiliation has some effect. Private sector means a relative swing to the right. All in all, however, the statistical effect of blue collar versus white collar is still stronger than sector affiliation.

Turning to post-war generations, we find that the stronghold of the radical wing is not blue-collar workers, but white-colfar workers affiliated with the public sector (see Fig. 3). Indeed, in the public sector, blue-collar workers are more inclined to support the conservative bloc than are whitecollar

Regional Trends: Voting patterns in Norway have traditionally been characterized
regional contrasts (Rokkan & Valen 1964). The long-term
trend however, has, been towards de-regionalization.

The different phases in the regional pattern of voting can be related to what Stein Rokkan has called 'the national revolution' and 'the industrial revolution'. The establishing period of Norway's political parties coincides with the national revolution, when regional contrasts were very clear.

The industrial revolution meant regional standardization. Industrial
plants were installed in different regions throughout the country. The
industrial worker became a transregional figure.

In Norway's first party election in the 1880s, during the period of the
national revolution, the centre-periphery profile was very clear-cut. This
conflict between centre and periphery was the focus of Stein Rokkan's

work. Rokkan discerned two peripheries, one in the south-western part of Norway - the cultural periphery - and one in the north - the class-polarized periphery (Rokkan 1967). Individuals in the cultural periphery channelled their claims through the countercultures which have the middle parties as their political arm. Consequently, the south-western periphery was dominated by the middle parties.

The class-polarized periphery in the north was a social and economic
protest which strengthened the socialist bloc. The north was a radical belt
in the political geography.

The centre in Rokkan's scheme consisted mainly of the capital, Oslo,
and its surroundings - traditionally a Conservative stronghold.

The Conservative breakthrough in the early 1980s meant a reinforcement

of the long-term trend of de-regionalization. The party's electoral base was enlarged to rural areas and a nationwide organizational network was built up. The centre-periphery profile in the pattern of voting was not as clear anymore, but it had not disappeared.

The old pattern can still be traced in the political geography. In Fig. 4 we have illustrated the relative strength of the various parties and party blocs from 1945 to 1987. In addition to the nationwide result, we have separate figures for the centre (Oslo) and the two peripheries (south-west and northern Norway).

In the south-western part, the middle parties have been in constant
decline since 1945. As regards party affiliation, the cultural periphery has
become similar to the national standard.

A process of equalization to a national norm can also be seen in the northern periphery up to 1979. But since 1979, the socialist bloc has increased its support, the northern part now seeming to be on the way towards regaining its position as a radical stronghold. From 1983 to 1987, however, the socialist bloc experienced a minor decline in the north. The decline for the radical wing was much stronger in the centre, in the capital, so the gap between the radical north on the one hand and the centre in the south has widened.¹⁴

To sum up, the regional contrast between east (Oslo) and south-west
seems to be gradually disappearing. However, the contrast between north
and the centre (or the capital) is being revitalized.

During the 1980s we can observe both a continuation of de-regionalization and the emergence of a regional conflict - the centre in the south versus the north - which we will relate to the transforming forces in post-industrial society.

This new regional conflict has been labelled 'sunbelt versus frostbelt' in the USA and 'south versus north' in Great Britain. The stronghold of the conservative wing is 'the sunbelt' in the USA and the 'south' in the UK. In Sweden, other terms have been used, 'growth Sweden' versus 'problem Sweden' (Holmberg & Gilljam 1987). In many ways both the Swedish and the British situations resemble the one in Norway. Northern Norway is a region with a declining population and a range of social problems.

The regional stronghold of the conservative bloc is the capital and more generally the major cities and their surroundings, i.e. the centre of growth in the country. 'Growth Norway', like 'growth Sweden', is a conservative stronghold.

The current regional contrasts in Norway are partly - but not exclusively -
a result of different occupational structures. In the north there are no large
cities, and the primary sector, dominated by fishing, is strong.

We have divided the different muncipalities according to the occupational
structures (see Table 5). Among the municipalities dominated by the

tertiary sector, the large cities (50,000 inhabitants or more) are sorted as a separate category.¹⁵ The five categories (primary sector, mixed primary/ secondary, secondary sector, tertiary sector, cities) in Table 5 can fall along a centre-periphery dimension going from rural to urban areas.

The two conservative parties have a clear-cut centre-periphery profile. The Party of Progress is three times as strong in the cities as in rural areas; the Conservative Party is twice as strong. From 1983 to 1987, the Conservative centre profile has been a little weaker. As the parentheses indicate, the decline in percentage points was strongest in the cities. The loss of votes for the Conservative Party is, however, more than compensated by the increase in votes for the Party of Progress. The point we want to stress is that the Conservative bloc grows stronger in the centre dominated by the tertiary sector. In the large cities, nearly every other voter supports the Conservative bloc, while in muncipalities dominated by the primary sector, only every fifth voter lines up behind the Conservative bloc.

In contrast, Labour is the weakest in the largest cities. This is a new phenomenon. In the post-war period up to the 1979 local elections, Labour always did better in the large cities than in the periphery dominated by the primary sector.¹⁶ From that point in time the situation has been reversed. In the election of 1987, Labour suffered a minor decline in the primarysector The long-term trend is, however, not only relatively, but also in absolute figures (compared for example with the elections of 1975 and 1977), a strengthening of its foothold in the periphery.

The most striking observation is that Labour's support does not follow any clear centre-periphery pattern. The Party's strength is in municipalities with a strong element of the secondary sector. The industrial worker is still the most loyal Labour voter.

Regarding the parties to the left of labour (Socialist Left/Communists) one can detect a systematic, but not strong, tendency towards increased support in the more central area. The middle parties, however, have a very clear-cut centre-periphery profile, two and a half as many voters in the periphery as in the centre (large cities). This party bloc is definitely related to the primary sector and the long-term trend of decline is partly a result of structural changes with farmers and fishermen as shrinking groups. The Conservative bloc, on the other hand, gains from a tertiary sector in growth.

In the centre, the volatility seems to be greater than in the periphery. The figures in parentheses, which show changes from 1983 to 1987, are an indication of this trend. High volatility means uncertainty regarding election results. The gains of the Conservative bloc in the centre do not have a firm foundation. An advantageous situation for the election's winner - the Party of Progress - is, however, that this party most clearly exposes the four cleavages which the article has stressed. The party is dominated by (a) men

in the (b) youngest generation settled in the (c) centre and often with
occupations not in public but (d) private sector.

Concluding Remarks

In one way the 1987 local elections were a deviant case: a protest election with a landslide victory to a recently established neo-liberalistic party, greater volatility among voters, and an extraordinarily low turnout. But, as we have noted, the election survey and other electoral statistics reveal a continuation and reinforcement of social and regional trends in voting patterns. These are trends which have emerged during recent elections, and are not unique to 1987.

Two cleavages have become more important: the gender gap and the division between voters working in the public and private sectors. Both these cleavages relate to the post-war generations, with older generations still tending to follow more traditional voting patterns. In addition, the regional contrast between the growth centre and the periphery has grown more significant. Thus, the 1987 'protest election' in fact confirms and reinforces familiar trends in overall voting patterns.

---

1. The turnout in the municipal council election was 69.4 percent. In the 1987 local elections there was also a ticket for the county council, and the turnout was 66.9 percent. Information about turnout among men and women is basedon turnout at either the municipal council or the county council election. Since a small percentage of the voters take part exclusively in the county council election, the figures for men and women are on a slightly higher level than the figures in the municipal council election (men: 70.9, women: 70.6). In order to have comparable figures with the elections before 1975 (when there were only municipal council elections), we have estimated the turnout for men and women in the municipal election. An assumption is that the proportions of men and women, among those who take part exclusively in the county council, are not biased with regard to sex.

NOTES

2. In Finland the Green party is represented in parliament by four MPs as a result of the 1987 parliamentary election. In Denmark the green alternative made itself felt for the first time in local elections in 1985 with a fairly good result, 2.8 percent, and with a couple of members in local councils. Sweden is the first Scandinavian country to establish a nationwide 'Green' party: this happened in 1981. The party grew out of the opposition against nuclear power which was organized as a forceful group during the referendum campaign on this issue. Although the party has never received more than around 2 percent of the vote in Sweden, much higher figures have been achieved in opinion polls between elections.

3. The election survey consists of two surveys, one from September and one from October. A question of voting behaviour in the county council election was posed. Data in Table 1 are weighed according to region, sex and party preference in the 1985 general election. Party distribution in the survey does not deviate much from the real election. For all parties except the Party of Progress, differences are under 2 percentage units, and are within the margins of sampling error. For the Party of Progress the difference is 2.7 percentage units compared with the county council election, and 0.5 with the municipal council election. Perhaps the respondents have confused the two forms of elections and thus under-represent the level of support in the county council election. At any rate, the deviances are far from dramatic, and I feel we can trust the data.

---

3. The election survey consists of two surveys, one from September and one from October. A question of voting behaviour in the county council election was posed. Data in Table 1 are weighed according to region, sex and party preference in the 1985 general election. Party distribution in the survey does not deviate much from the real election. For all parties except the Party of Progress, differences are under 2 percentage units, and are within the margins of sampling error. For the Party of Progress the difference is 2.7 percentage units compared with the county council election, and 0.5 with the municipal council election. Perhaps the respondents have confused the two forms of elections and thus under-represent the level of support in the county council election. At any rate, the deviances are far from dramatic, and I feel we can trust the data.

4. This governmental crisis had its background in the fall in oil prices, an important factor for an oil-producing country like Norway. Economic restrictions were necessary. The Party of Progress opposed new taxes, and Labour disagreed with the economic policy pursued by the Willoch government, claiming that this policy would increase income differences.

5. On the opinion polis conducted by MMI the Party of Progress increased its support from March to May from 2.8 to 5.5 percent.

6. The Norwegian case is very similar to that in Denmark. In Sweden, however, the proportion of immigrants from the Third World is much higher.

7. In Denmark the number of applicants in 1985 and 1986 was as high as the 1987 Norwegian level, but then it sharply declined as a consequence of a more restrictive pratice (Kørmendi 1987). The 1985 Danish situation was in many ways similar to that experienced in Norway in 1987: a sudden explosion in the number of asylum-seekers leading to unrest on the grassroots level (cf. Kalundborg).

8. The total base for the mock elections in 106,183 pupils from 413 secondary schools. Information about party preference for boys and girls separately is only available from 33 schools or from 4,651 girls and 4,141 boys. The sum of the votes from girls and boys does not deviate much from the party distribution from the total base of pupils. In other words, I think the gender gap is representative for the whole sample.

9. Professor Henry Valen and research fellow Bernt Olav Aardal at the Institute for Social Research are in charge of the Norwegian Electoral Programme.

10. Deviances from the 1987 election result were 2.0 percentage units or less. Turnout however, was overestimated: 19.6 percent non-voters, versus 33.6 percent as the correct figure. In relative figures, turnout was even more overestimated in the 1985 survey: only 7.5 percent indicated as non-voters, versus 15.9 as the correct figure. The proportion of non-voters increased by 12.1 percentage units in the panel compared to 14.0 percentage units in the electorate. We may thus conclude that the result in the panel fits quite well with the actual distribution. In the sample distribution of party strength in the 1985 survey there is, however, a significant statistical deviation from the actual election result. Labour was supported by 40.8 percent in the sample, but only by 36.9 among the actual voters. This deviation is a reminder that Table 3 does not show the exact figures in the electorate, although the overall tendencies are quite trustworthy.

11. A complete cohort analysis presupposes time-series data. In our analysis we have data from only one point of time. In faet the different generations have to be regarded as labels for different age groups.

12. The percentages in Figure 2 are basedon the following number of observations: Men Women 1961-69 116 112 1946-60 233 230 1926-45 227 226 -1925 185 235

---

13. The percentages in Figure 3 are basedon the following number of observations: Born after 1945 Bom before 1945 Blue Collar White Collar Blue Collar White Collar Public 126 82 88 103 Private 196 143 76 62

14. In Tables A and B these tendencies are documented. The central figures are the percentage differences. In Table A these figures are from 1965 to 1987 increasing concerning the socialist bloc, whereas Table B shows the opposite tendency. Table A. Elections results. Northern Norway vs. Oslo. 1965 1975 1979 1983 1987 Socialist North 58.0 50.8 47.8 56.0 54.0 bloc Oslo 49.5 45.3 41.4 44.6 41.0 8.5 5.5 6.4 11.4 13.0 Middle North 24.6 30.0 23.7 17.9 17.4 parties Oslo 11.2 12.2 10.6 8.6 8.5 13.4 17.8 13.1 9.3 8.9 Conservative North 17.3 18.4 27.0 25.4 28.6 bloc Oslo 38.6 39.5 47.9 46.7 50.5 -21.3 -21.1 -20.9 -21.3 -21.9 Table B. Elections results. South-western Norway vs. Oslo. 1965 1975 1979 1983 1987 Socialist South-west 38.8 32.3 30.7 36.0 34.6 bloc Oslo 49.5 45.3 41.4 44.6 41.0 -10.7 -13.0 -10.7 -8.6 -6.4 Middle South-west 44.3 42.4 35.7 30.9 31.2 parties Oslo 11.2 12.2 10.6 8.6 8.5 33.1 30.2 25.1 22.3 22.7 Conservative South-west 17.1 22.4 32.0 32.2 34.2 bloc Oslo 38.6 39.5 47.9 46.7 50.5 -21.5 -17.1 -15.9 -14.5 -16.3 Source: Official electoral statistics from Norwegian Social Science Data Service.

15. Municipalities with more than 50,000 inhabitants include the six largest cities - Oslo, Bergen, Trondheim, Stavanger, Kristiansand, Drammen - and Bærum, which can be regarded as a suburban municipality to Oslo.

16. This observation is basedon the whole country, large regional contrasts in the pattern of voting being disregarded.

Acknowledgment. - Stig P. Pettersen prepared all the graphics. I am grateful for his excellent
assistance.

REFERENCES

Bjørklund, T. 1986. 'Kvinners og menns partipreferanse', Tidsskrift for samfunnsforskning,
27, 417-443.

Burklin, W. P. 1987. 'Governing Left Parties Frustrating the Radical Non-established Left:
the Rise and Inevitable Decline of the Greens', European Sociological Review, 2,109-126.

Duverger, M. 1955. The Political Role of Women. Paris: UNESCO.

Holmberg, S. & Gill jam, M. 1987. Våljare och val i Sverige. Stockholm: Bonniers.

Kørmendi, E. 1987. Refugees in Denmark. The Danish National Institute of Social Research,
Booklet 24.

Listhaug, 0., Miller, A. H. & Valen, H. 1985. 'The Gender Gap in Norwegian Voting
Behaviour', Scandinavian Political Studies, 8, 187-206.

Mannheim. 1972. 'The Problem of Generations'. In P. Kecskemeti, ed., Essays on the
Sociology of Knowledge. London: Routledge & Kegan Paul.

Rokkan, S. 1967. 'Geography, Religion and Social Class: Cross-Cutting Cleavages in
Norwegian Polities'. In Lipset, S. M. & S. Rokkan, eds., Party Systems and Voter Alignments,
York.

Rokkan, S. & Valen, H. 1964. 'Regional Contrasts in Norwegian Polities'. In Allardt, E. &
Littunen, Y., eds., Cleavages, Ideologies and Party Systems. Helsinki: Westermarck Society.

Valen, H. 1976. 'Norway: The Local Elections of September 1975', Scandinavian Political
Studies, 2, 168-183.

Valen, H. & Aardal, B. 1983. Et valg i perspektiv. En studie av Stortingsvalget 1981. Oslo:
Statitisk Sentralbyrå.