Norway experienced her first regular "ins-outs" election in September 1969: the old "in" party, Labour, had been in opposition since 1965 and made a valiant effort to gain the majority required for a return to executive power.¹ The fight was very close: for a while during the exciting election night the computers in fact seemed confident that Labour would make it back to the majority position. The final result was as close to a deadlock as it is possible to get in the Norwegian Parliament: Labour increased its strength from 68 to 74 seats, while the governing coalition of four non-socialist parties scraped through with a minimal majority of 76. In fact the vote was so close that a transfer of a few hundred votes could have evened the score between the two blocks: in the Troms constituency Labour came within 900 votes of winning an additional seat.

The dramatic gains for the Norwegian Labour party parallel several similar Social Democratic surges toward the majority point in Western Europe. In Sweden the Social Democrats made a spectacular comeback in September 1968: the party had fallen as low as 42.2 % at the local elections in 1966 but got back to 50.1 % at the national election of 1968.² In the German Federal Republic, the Social Democrats finally broke through in September 1969 and gained the right to take the lead in the formation of a kleine Koalition. And in Austria the elections of March 1970 brought the SPD back to prominence after four years in opposition: even in this Proporz democracy elections at last seemed to offer a chance for some alternation in power. In all these cases the Social Democrats have had to face difficult issues of long-term as well as short-term strategy: should they "go it alone" and work toward majority power all on their own or should they negotiate some alliance with one of the possible partners to the left or to the right? In Sweden, the Social Democrats were able to reach the majority point on their own in 1968 but much of the debate among the leaders over the changeover to a unicameral Riksdag and over the new electoral law hinged on considerations of alternative coalitions or exclusions.³ In Denmark, the Social Democrats had traditionally sought their alliances on the right with the Radicals, but for a brief period found it more rewarding to rely on an informal arrangement for support with the left-wing Socialist People's party: this, however, quickly

came to an end after the split-up of the ally.4 In Germany, the Social Democrats were faced with two alliance options, both to their right: a kleine coalition with the Liberals, a grosse with the Christian Democrats. Willy Brandt courageously chose the harder course and worked out an alliance with a weakened but radicalized FDP. In Austria, Bruno Kreisky was left with no such choice: a coalition with the Nationalist-conservative Freedom party would clearly be suicidal; the only realistic options were between a return to the old and well-tried formula of grosse Koalition and Proporz government, and a desperate attempt at minority parlamentarism.

In Norway the near-victory of the Labour party brought out into the open a number of questions about the ranges of options in the inherited system of constellations: should Labour stick to its "splendid isolation" and refuse to contemplate any coalition or should the party respond in some positive way to overtures from such possible partners as the radicalized wing of the Liberal party or even, what remained much more speculative, from the old ally of the 19305, the Agrarian Centre party? The heavy losses suffered by the Liberal Venstre, particularly in its Western strongholds, triggered a crisis of confidence in the party leadership and generated widespread speculation about the potentialities of a Brandt-Scheel type coalition in Norway. The situation was no less ambiguous on the left front of the Labour party: the Left Socialist party was deeply demoralized after the defeat and a number of the local leaders withdrew from the party and offered to support Labour in the municipal councils.

For all these reasons the election of 7—8 September 1969 opened up a new phase in the development of the Norwegian party system. To gain some understanding of the changes in the conditions for parliamentary bargaining, we shall clearly have to analyse a wide variety of data on the failures and successes of the party campaigns in the different sections and strata of the expanding electorate. In any such analysis it will be essential to go back to the stable period through 1957 and to follow the process of change step by step: from the initial defeat of the Labour party in 1961 through the mobilization of new support at the local elections of 1963 and the crushing defeat in 1965⁵ to the nearmiss in 1969. Happily we have been able to build up a solid basis for analyses of this kind within our joint Oslo—Bergen programme of electoral research: we have not only assembled a wide variety of ecological and organizational information commune by commune,⁶ we have also brought together a considerable file of biographical information on candidates and representatives, and have carried out cross-sectional panel surveys of the electorate in 1965 and 1969.⁷ In this brief note we shall confine ourselves to only very elementary tabulations, however: fuller accounts of our analytical efforts will be published during 1971.

We shall first present the overall statistics for 1965 and 1969 and point out differences in the results for the parties from one election to another. We shall next present a simplified transition matrix from our panel survey and discuss some crude evidence for hypotheses about the major sources of instability.

The resurgence of regional contrasts

Table 1 gives the nationwide results for each party at the elections of 1965 and
1969.

Labour stands out as the definite winner of the election: the party gained more percentage points than any of the other parties and was even more rewarded at the allocation of seats. The table leaves little doubt that a substantial number of the votes gained must have been cast by supporters of the splinter parties to the left: the Socialist People's party lost 2.5 percentage points and the Communists another 0.4 points while Labour gained 3.3. The Labour party centred its entire campaign on the question of government power: the election was presented as a fight between two alternative teams of governors, the "ins" versus the "outs". This clearly netted the party a number of new votes: a considerable proportion of the left Socialists were persuaded that the election was not an occasion for the expression of intransigent counter-identities but an opportunity to change their governors. The deep ideological split within the Socialist People's party no doubt contributed further to the strength of the Labour party: the uhra-gauchiste student faction which left the party in the winter of 1968—69 in fact adopted the French slogan élection-trahison and condemned parliamentary politics as a sham and a waste of time. But the Labour party must have done better than just rally some of its dissidents: the total vote for the three-party socialist block increased by 0.6 percentage points to a clear majority of 51.1 °/o. But this majority of votes did not produce a majority of seats: all the votes of the two splinter parties were wasted below the threshold of representation.8 In 1965 the CP had presented lists in 16 of the 20 constituencies without gaining a single seat. The Socialist People's party had presented lists in all the 20 constituencies and won in only two: in the capital and in the second largest constituency, Nordland. In that election altogether 11.3 °/o of the socialist total was wasted below the threshold: this compared with 6.9 % for the non-socialist block.

This gave the four parties on the right a definite advantage: they gained most of their seats at the cheap first-seat level and were able to achieve representation at a considerably lower cost than the Socialists. In 1969 the balance tilted the other way: the two Socialist splinters lost votes and reduced the percentage below the threshold for their block to 8.6, while the non-Socialist parties wasted many more votes than in 1965 and in fact reached a higher below-threshold percentage than the Socialists.

The electoral system was less friendly to the non-Socialists in 1969 than it had been in 1965: the victory in 1965 was achieved through an over-representation for the entire block of 3.8 points; in 1969 this was reduced to 1.8. To understand how this came about we shall have to look at the results for each of the constituencies (Table 2).

In 1965 the Socialist block presented a total of 56 lists: 34 of these went without representation, three of them, the two Left Socialist ones and one Labour, won one seat only while the remaining 19 all gained two or more seats each. The four non-Socialist parties had a total of 71 lists in the field: of these only 14 went without representation while as many as 37 won only one mandate. In 1969 there were 58 Socialist lists but only the 20 Labour lists gained representation: all at the 2 + level. The non-Socialist lists fared worse: there were 73 lits alltogether, but the number of below-threshold lists increased from 14 to 19. Three new lists proved unsuccessful: the ones presented by the Liberals in Østfold, the Christians in Hedmark and the Agrarians in Finnmark.

Only one of the below-threshold lists from 1965 was successful in 1969: the Christian list in Oslo. Of the one-seat lists from 1965, the Liberals proved the most vulnerable: two of them lost to Labour (Vestfold and Sogn and Fjordane), a third to the Agrarians (Sør-Trøndelag). The Conservatives unexpectedly lost their one seat in the inland constituency of Oppland: this reduced the representation in that district to two parties, Labour and the Agrarians. All the other non losses were at the two-seat level: the Conservatives lost their second seat in Østfold while the Liberals lost theirs in the heart of the West country, in Rogaland and Hordaland. These losses at the two-seat level all represented gains for competitors well beyond the threshold: to three- or four-seat Labour lists. Interestingly enough, only one of the losses suffered in the 1969 election was made at the margin: only in Vestfold did the last-seat party from 1965 lose out in 1969. This Liberal loss must be understood against the changes in the list structure in the constituency: the Christians had been kept out for tactical reasons in 1965 but returned in 1969 on a joint list with the Agrarians; this clearly also affected the chances of the Liberal list. In all the other cases the losers had had reason to feel at a safe distance from the margin after 1965. This suggests that there was some pay-off from marginality: it seems to have been easier to mobilize new voters and even tactical cross-voting when the fate of the seat was clearly in jeopardy. We do not know enough about the extent of such cross-party voting in defence of marginal seats but hope to analyse further evidence in later reports.

The changes from 1965 to 1969 again underscored the importance of the

inherited regional contrasts in Norwegian politics.⁹ These contrasts come out particularly clearly in the distribution of votes and are only faintly reflected in the change-overs of seats. The Labour party fared particularly well in the North: if we compare the Left Socialist losses with the Labour gains we find an average surplus of 6 percentage points in the three constituencies up north. These spectacular gains in votes found no reflection in the distribution of seats, however: Labour had too many seats already and the Sainte—Lagiie formula enforces increasing costs for each additional seat. The pay-off for increases in voting strength was greatest in the West: there the Labour party had still a far way to go to reach the majority of seats; the cost of additional seats was consequently lower. Three of the Labour gains from the Liberals occurred in the constituencies with the lowest shares of Labour seats: Rogaland (1965: 3 Lab. out of 10), Hordaland (also 3 out of 10: a very close race in 1969) and Sogn and Fjordane (1 out of 5 in 1965). The other Labour gains were made in the East: two of them brought about through heavy Conservative losses in competition with Christians and Agrarians (Østfold and Oppland), the third in a constituency affected by changes in the list structure and extensive cross-voting in the non-Socialist camp (Vestfold).

Even more important for the future of the party system were the changes brought about in the balance between the three parties in the middle of the political spectrum. The Liberal losses were markedly heavier In the southwest and the west than in the rest of the country but these losses were, with great regularity, balanced off against corresponding gains for the Christians. The antifundamentalist policy postures of the Oslo—Bergen dominated Liberal leadership had split the party in its traditional strongholds in the southwest and west and motivated at least some loyal Venstre supporters to vote for the Christians in 1969: the frequent radical attacks against the religious and the mass-media policies of the Christian leader, the Minister of Education, Kjell Bondevik, seemed to have spurred a strong mobilization drive among the fundamentalist Lutherans in the counter-cultural south and west. The consequent losses in Liberal seats changed the regional balance in the middle of the political spectrum:

The Liberals had suffered their heaviest losses in their ancient heartland and no longer dominated the anti-socialist front in that region. The Christian People's party had taken over the lead in the Southwest/West and had, in the process, become much more regionally distinctive in its profile: 9 out of its 14 representatives

were elected from that region, only 6 out of 13 Liberals. The Agrarians had consolidated their strength in the eastern valleys and in the Trøndelag: they ran neck and neck with the Conservatives in the Inland East and they dominated the government block in the home province of the Prime Minister, the Trøndelag. The Conservatives remained the leader of the non-Socialist camp in the central areas around the capital but even there suffered uncomfortable losses. In the far north the Labour party held a clear majority while the government votes were more evenly spread than in any other region. The Labour party had clearly been able to mobilize a much greater number of peripheral protest votes than the Left Socialists in 1961 or 1965: for a large party it really pays to be in opposition for a while! The Labour gains were spectacular even in the nothernmost of the constituencies, in the traditional stronghold of the anti-system left. In Finnmark the CP maintained a remarkable hold on its clientele (4.2 °/o in 1965, 4.1 % in 1969) and the SP lost much less than elsewhere (the party collected 6.5 % of the votes in 1965 and 5.8 in 1969: if we set the 1965 share at 100, the national average fell to 58 while the Finnmark score stayed at 89). Even so, the Labour party did much better than average in this northern periphery: the nationwide increase was 3.3 points, in Finnmark it was 5.7.

Further ecological analysis will be required for any precise mapping of these increases in the regional distinctiveness of the distributions of the votes among the parties. These constituency-by-constituency tabulations bring out clear contrasts and suggest hypotheses for further testing. We shall briefly explore two such hypotheses in this preliminary analysis: one concerning the transition of votes from the Socialist People's party to Labour, the other one concerning changes from the Liberal party to the Christian People's party.

In Table 3 the communes have been classified according to the stregnth of
the Socialist People's party in 1965.

As might be expected, the commune-by-commune ecological analysis supports our initial observation that the gains of the Labour party and the losses of the Socialist People's party tend to cancel each other out. Consequently, it is not surprising that Labour in 1969 enjoyed the greatest gains in areas where the Socialist People's party held a relatively strong position at the preceding election. However, an interesting difference appears when we compare communes by level of urbanization. In the least urbanized communes, the sparsely populated fishery communes, the gains of the Labour party as well as the losses of the Socialist People's party were greater than in more urbanized areas. In sparsely populated communes, particularly fishery communes, the gains of the Labour party exceeded the losses of the Communists and the Socialist People's party. In densely populated rural communes, as well as in urban areas, changes in vote proportions between the Socialist parties came close to a balance.

Next we turn to the question of vote exchanges between the Christian
People's and Liberal parties. In Table 4 the communes have been classified according
to the strength of the Christian People's party in 1965.

Table 4 clearly demonstrates that the stronger the position of the Christian People's party was, the greater were the losses of the Liberals. This tendency holds both for sparsely and densely populated communes. And the same regularity is apparent in all regions, but less in the East than in other parts of the country. The support for the Christian People's party may be taken as an indicator of the strength of the lay religious movements.

Table 4 supports our initial hypothesis that the Liberal party suffered its greatest losses in areas where religious sentiments are strong. Apparently a large number of those who left the Liberals in 1969 went to the Christian People's party. But if we compare the changes commune by commune, there is no clear consistency between the losses of the Liberals and the gains of the Christian People's party. Table 4 suggests that the defectors from the Liberal party must have moved in several directions.

Commune-by-commune ecology will help us to pinpoint the areas of maximal and minimal change and suggest propositions about conditions for substantial transfers of votes within the electorate. But the commune figures can never become more than net aggregates: to test hypotheses about transitions from one election to another we shall clearly have to link up ecological-level data with individual-level data from surveys. This requires complex analysis designs: in this brief note we can only give a couple of elementary tables from our panel survey.

The movement of voters across party lines

As David Butler and Donald Stokes have demonstrated with such clarity in their volume on the recent elections in Britain,¹⁰ the net changes in party strengths from one election to another are functions of in-movements and outmovements at three levels:

— at the level of the electorate: the intake of young entrants and naturalized
citizens must be balanced against the losses through death and emigration;

— at the level of participant voters: the numbers mobilized from the stratum
of abstainers must be judged against the numbers of backsliders into abstention;

— at the level of each party: the votes gained from other parties must be
balanced against those lost through changes of preference.11

In Norway there was a marked increase in the total electorate from 1965 to 1969: the age threshold for entry was lowered from 21 to 20. From 1961 to 1965 the electorate increased by 2.8 °/o, from 1965 to 1969 by 7.2 °/o. This marked increase was not only due to the lowering of the franchise: 1969 also saw the entry of the "bulge" generation into national politics. The Central Bureau of Statistics has estimated that the total number of new entrants was 316,000: roughly 12 °/o of the 1969 electorate. On the other hand the losses through death or emigration were roughly 143,000: this explains the net increase of 7.2 °/o.

Table 5 tells us about the consequences of this increase. It has generally

proved difficult to achieve adequate representation of the youngest members of the electorate in sample surveys but the differences we can report from 1965 to 1969 certainly fit in with the other evidence we have at hand: Labour gained back some of the confidence it had lost among the young in 1965 while the Left Socialists proved unable to make use of the great opportunities offered through the lowering of the franchise to the 20-year-olds.

The differences in turnout level were less important: the national average was 85.4 °/o in 1965, 83.8 °/o in 1969. Most of this decline must clearly be attributed to the increase in the proportion of first-time voters: the low turnout figures for new entrants into the electorate have been repeatedly documented.¹² Our data from panel interviews in 1965 and 1969 make it possible for us to estimate some of the consequences of fresh mobilization vs. abstention, but this analysis is fraught with great difficulty. We were able to check turnout for the panel respondents against the local registers for 1965 and we also had them report nonparticipation in the 1969 interview but both percentages are much lower than those reported in the official statistics: the panel clearly over-represents the stable participants in the electorate.

Clearly, to get better estimates of the consequences of movements in and out of the core of registered voters, we shall have to analyse in some detail the respondents who could not be interviewed twice: we have already reported a set of data for those who were on the register for the first time in 1969 but we shall also have to have a good look at the turnout record and the reported votes of the others who could only be interviewed once.

In this round, however, we shall confine ourselves to the 1965—69 panel. We were able to interview 1079 respondents in our nation-wide cross-section on two occasions. Of these, 32 refused to indicate either their 1965 vote or their 1969 vote. Another 50 were checked out as non-_voters on the registers in 1965. Another 36 said they did not vote in 1969. This leaves only 961 twice-voters

with known party preferences. In this analysis we shall include in our tables the 86 who failed to vote at one or both of the elections: this will allow some discussion at least of the consequences of movements in and out of the ranks of the participant voters,

Table 6 gives the full matrix for the 1047 cases. Table 7 concentrates on the movement between the Left Socialists and Labour within major educational strata: this was the movement that contributed most to the over-all change between the two elections and therefore deserves special attention.

As so often before, Labour and the Agrarians, the two parties with the firmest
basis in the social structure, stand out as the ones with the largest proportion
of loyal voters: on this score there is no difference between educational levels.

The Christians and the Conservatives have smaller cores of stable voters and in these parties education clearly matters: particularly in the case of the Conservative party there is a marked difference in the level of stability between the less educated and the more educated. The parties subject to the greatest amount of change were the Liberals and the Left Socialists. In both cases education made some difference. The Liberals retained barely half of their 1965 voters: they lost heavily to Labour in the lower educational strata; in the higher more often to the Agrarians and the Conservatives. The direct losses to the Christians seem to have been less important: this is surprising in the light of the evidence in Table 2 of the fit between the net losses of the Liberals with the gains for the Christians in the Western constituencies. This may be brought out more clearly in regional breakdowns. As suggested in Table 4, the net Liberal losses were brought about through complex movements across three or four party fronts: the party made considerable gains from the Conservatives in many districts but these gains were offset by greater losses to Labour and the Agrarians, while all these parties in their turn lost some of their votes to the Christians. These movements cannot be explored without further controls by region and by level of urbanization: the difficulty with any such controls is that the N for the Christians is uncomfortably low for both elections in the panel. Here again comparisons with the respondents who could only be interviewed once may pay off.

The most striking finding in these transition matrices is for the transfers of votes from the Left Socialists to Labour: practically half of those who voted for the Left splinter in 1965 voted for Labour in 1969. The educational control offers clear evidence that it was the traditional working class voters, those with least education, who turned back to the fold: of those with only primary education as many as 56 % voted Labour in 1969; of those with further education, only 42 °/o. This fits in with other evidence for the Socialist People's party: the educational levels of the SP voters were markedly higher than those of the Labour voters in the central regions of the country, while they were much more similar in the western and the northern peripheries. Ottar Hellevik, using Gallup data for 1965, has characterized the Socialist People's party as an uneasy alliance between an urban pacifist-intellectual movement and an agency for peripheral protest: it was no accident that the party won one seat in the intellectual centre of the country, the other in a northern periphery.¹³ The results for 1969 suggest that the party suffered its heaviest losses in the social and the geographical peripheries: the losses among the least educated are as marked as the losses in the sparsely populated communes, particularly fishery communes in the two northern constituencies, Nordland and Troms.

These early findings suggest important items for further analysis and fresh data-gathering. The 1969 election brought into the national electorate the first products of the educational explosion of the 1960'5, but the immediate impact was one of stabilization: the Labour party made a spectacular comeback and the regional differentiations within the non-Socialist block reasserted themselves

with great clarity. Does this mean that the party leaders will find ways of controlling even this great source of disequilibrium? The current party system has already survived two generations of continuous social change: the decline of agriculture, the rapid increases in urbanization and the spread of urban life styles, the growth of the services and the public sector. Will the system even survive the fundamental changes brought about through the violent increases in levels of scholling and in exposure to information? This will no doubt be a central theme in any research on mass politics in the 1970'5.

Henry
Universities of Bergen and Oslo

Stein Rokkan and Henry Valen

---

NOTES

1 For a discussion of the position of the Norwegian party system at some point half-way between "ins-outs" alternation and multiparty coalition bargaining, see Stein Rokkan. "Norway: Numerical Democracy and Corporate Pluralism" in R. A. Dahl (ed.). Political Oppositions in Western Democracies, New Haven, Yale Univ. Press, 1966. For further refinement, see Olof Ruin. "Patterns of Government Composition in Multi-Party Systems: The Case of Sweden", Scandinavian Political Studies, vol. IV, 1969, pp. 71_—87.

2 On the Swedish election of 1968 see Bo Sarlvik, Valet 1968 (Gothenburg: Institute of Political Science, 1969 mimeo); to this is appended an English-language explantory note "The 1968 Election in Sweeden". See also his contribution to this volume of SPS, p. 241 ff.

3 See Olof Ruin, op. cit.

4 See L Glans. "The Danish Parliamentary Election of 1966", Scandinavian Political Studies, vol. 11, 1967, pp. 266^—272, and E. Damgaard. "The Parliamentary Basis of Danish Governments", Scandinavian Political Studies, vol. IV, 1969, pp. 30^—57.

5 For analyses of the 1965 election, see S. Rokkan and T. Hjellum. "The Storting Election of September 1965", Scandinavian Political Studies, vol. I, 1966, pp. 237^—246; H. Valen. "Partiforskyvninger ved stortingsvalget i 1965", Tidskrift for samfunnsforskning, vol. VIII, 1967, pp. 113_—142; H. Valen and P. Torsvik. "Økningen i valgdeltakelsen ved kommunevalget i 1963 og stortingsvalget i 1965", Tidskrift for samfunnsforskning, vol. VIII, 1967, pp. 187^—218.

6 See S. Rokkan and F. Aarebrot. "The Norwegian Archive of Historical Ecological Data", Social Science Information, vol. VIII, no. 1, 1969.

7 For general accounts of the Norwegian programme of electoral research, see S. Rokkan. Scandinavian Political Studies, vol. I, 1966, pp. 266_—280, and H. Valen and S. Rokkan. Scandinavian Political Studies, vol. 11, 1967, pp. 294^—305. Major publications under the programme include H. Valen and D. Katz. Political Parties in Norway, Oslo, Universitetsforlaget, 1964, and S. Rokkan et al. Citizens, Elections, Parties, Oslo, Universitetsforlaget, and New York, D. McKay, 1970.

8 For further details on the working of the electoral system, see S. Rokkan and T. Hjellum, op. cit.

9 For a first account of these differences, see S. Rokkan and H. Valen. "Regional Contrasts in Norwegian Politics", pp. 190^—247 in E. Allardt and S. Rokkan (eds.). Mass Politics, New York, Free Press, 1970. A few further details are given in S. Rokkan. "Geography, Religion and Social Class", pp. 367_—444, in S. M. Lipset and S. Rokkan (eds.). Party Systems and Voter Alignments,New York, Free Press, 1967.

---

10 D. Butler and D. Stokes. Political Change in Britain, London, Macmillan, 1969.

11 For a detailed discussion of the implications of these three sets of movements for the evaluation of the fit between survey data and official statistics, see S. Rokkan. "Velgerreaksjoner og valgsprognoser" in B. Alstad (ed.). Norske Meninger, Oslo, Pax, 1969, vol. 11, pp. 11^—26.

12 See especially S. Rokkan. Valgdeltakelsen blant de yngste, Bergen, Michelsen Institute, 1964. For details on the high turnout level in 1965, see Valen and Torsvik op. cit.

13 See O. Hellevik. "Noen trekk ved den norske SF-velger", Kontrast, no. 10, 1968.