1. Introduction

Some years ago prospective students of politics were taught that the number of political parties was a significant characteristic of political systems; it was particularly important whether a country had a two-party or a multi-party system.¹ The prototypes were, on the one hand, the stable, efficient, and 'good' two-party systems of the United States and Britain and the unstable, inefficient, and rather 'bad' systems of pre-1958 France, Italy, and Weimar Germany, on the other.²

The two/multi-party line of thought was dominant to a degree that gave Dankwart Rustow good reasons for writing an article entitled 'Scandinavia: Working Multiparty Systems' to highlight his point that 'Scandinavian experience belies the frequent generalization that political stability can result only from a two-party system'.³ Other authors were equally dissatisfied with the dichotomous classification of party systems. Giovanni Sartori, for example, found that one could not deal with multi-party systems as a single category, since there was a world of a difference among some of the systems falling into this category. Instead, he proposed to distinguish among 'simple pluralism' (e.g. Britain), 'moderate pluralism' (e.g. the Scandinavian countries), and 'extreme pluralism' (e.g. Italy), concentrating himself mainly on the latter type of party systems.⁴ Arend Lijphart, to give another example, developed a model of 'consociational democracy' to explain the stability of the multi-party systems of the Netherlands, Switzerland, and Austria. The Anglo-American and the Scandinavian systems, among others, were referred to as 'centripetal democracies', while Italy, France, and Weimar Germany were called 'centrifugal democracies'.⁵

---

* This paper was written a few weeks before the December 1973 election was suddenly called. Since it deals with the structural characteristics and long-term development of the party system since the 19205, I decided to leave the manuscript as it was and to add a short section on that highly unusual election of 1973.

It thus appears that, once the smaller European democracies were included in analyses, there was general agreement that one could indeed have a multi-party system along with political stability.⁶ The aim of this paper is not to propose another theory or typology of democratic systems, but, more specific and concrete, to analyze - at the mass level as well as the elite level - the development of the Danish party system in the past half century as it relates to political stability and change. First, I shall present data on voter support and parliamentary strength of political parties within a longitudinal perspective. Secondly, various forms of parliamentary party behavior will be analyzed to discover sources of legislative stability. Thirdly, I shall deal with aspects of government stability and performance. Next, special attention will be paid to the major changes that occurred in the party system during the 19605. Finally, I have added a short review of the remarkable December 1973 election.

2. The Danish Parties Through 50 Years: Electoral Support and Parliamentary Seats

Lipset and Rokkan tell us that 'the party systems of the 1960s reflect, with few but significant exceptions, the cleavage structure of the 19205'.⁷ They list 'four critical lines of cleavage': (1) subject versus dominant culture; (2) church(es) versus government; (3) primary versus secondary economy; (4) workers versus employer, owners.⁸ Other lines of cleavage may have been present to some extent, but the basic cleavage structure of Danish politics in the 1920s consisted of cleavages 3 and 4. The Social Democrats was the party of the workers, the Liberal party {Venstre, sometimes translated into Agrarians or Agrarian-Liberals) the party of the farmers, and the Conservatives the party of employers and upper civil servants; the Radical Liberals did not as easily fit with a distinct social grouping, although they received strong backing from the smallholders. In addition, however, the Radical Liberals attracted support from a wider range of citizens than any other party.

All of these four parties could trace their history back into the previous century. In the 19205, with the big constitutional issues behind them (cabinet responsibility, universal suffrage, proportional representation (PR)), the four 'old' parties were located in the same order (SD, RL, L, C) on a left-right axis on most⁹ important policy issues of the times like, e.g., social welfare legislation and defense policy.¹⁰

Thus, in 1920 the stage was set for a new era of parliamentary democracy in Denmark. At the September 1920 election to the Folketing, the four old parties polled a total of 96.2 % of the national vote. My first question is: What happened to their dominant position in the following 50 years? From 1920 on, PR gave new parties a better chance to break through the electoral barrier than the earlier firstpast-the-post system. In fact, eighteen different new parties have participated in Folketing elections since 1920, with an average of four to five new parties at each election. The election results are shown in Table I.

In Fig. 1 the combined vote of the four old parties is graphed for each of the nineteen elections. The answer to the question raised is that the old parties have kept a dominant position in terms of voter support, but with a long-term trend of some decline. We can get a more exact answer by grouping the nineteen elections in four time periods, each containing four or five elections, and taking the average percent of voter support for the old parties, cf. Table 11. We find then that the

combined electoral strength of the four parties has declined by about 11 % from
the first to the last decade included in the analysis.

There are, of course, some fluctuations around this long-term trend of decline. On the basis of Fig. 1 and Table I we can make the following observations: in the 1920s the four old parties were virtually totally dominant, whereas the 1930s showed some erosion of support due to the emergence of several small new parties. In the atypical election of 1943 (during the German occupation of the country, when the Communists were outlawed and with the German minority party not running for office) a temporary recovery took place, only to be followed, in 1945, by the lowest percentage in the whole century, owing to a strong showing of the Communist party. The fifties, again, was a period of stability, but with a lower level of support than in the 19205. Finally, in the sixties, new oscillations occurred, and vote shares were obtained by the new Socialist People's party in particular.

Since the percentages include the votes obtained by four different parties, we might want to know how the long-term development of each party's voter support relates to the overall decline. Table I shows that in the sixties the Social Democrats and the Conservatives had roughly the same percent of the vote as in the twenties.

The Radical Liberals steadily lost ground until the end of the sixties, but then the party recovered and even obtained slightly more votes than in the twenties. Consequently, the overall decline of the four-party percent of the vote is completely explained by shrinking support for the Liberal party. An obvious reason for the decline is the narrowing of the party's traditional supporting group of farmers.

Thus far I have briefly outlined the long-term development of voter support for the four old parties. We must also look at the parliamentary level, however, since the parliamentary party structure is not a totally accurate reflection of the party structure at the mass level. Even if votes are translated into parliamentary seats by the PR election system, some distortion is bound to occur because of the electoral thresholds¹¹ that parties have to pass to obtain seats at all. In Fig. 1I have graphed the percentage of votes versus the percentage of seats for the four old parties. The figure clearly shows that the old parties generally have received more seats than strict proportionality dictates. Their combined share of the seats has decreased, however, from 97.9 % in the twenties to 89.8 % on the average in the latest five elections. But this decline of 8.1 % is less than the 10.8 % vote share reduction shown in Table 11, which means that electoral thresholds to some extent have contributed directly to preservation of old party preponderance. The indirect effect of the sheer existence of electoral thresholds on new party formation and voter support cannot be measured, of course, but it might be more important than the direct effect.

By returning once again to Fig. 1, we can elaborate a little further the point of the discrepancy between seats and votes. The lower half of the figure shows the percent of votes obtained by new parties that managed to surmount electoral thresholds and by new parties that did not ('voters not represented'). Until 1947, and especially from 1932 to 1947, the old parties¹² received additional seats at the expense of numerous small parties that were able to capture only a few seats. In 1953, a coalition of the three largest parties (SD, L, C) decided to raise the threshold resulting in an increasing number of voters not being represented by their preferred new party (in 1971 no less than 6.8 %, corresponding to eleven of 175 seats) and in 'overrepresentation' of the major parties. In a comparative perspective, the overrepresentation of major parties may appear insignificant. However, with four or five major parties in Parliament, coalition politics is seriously affected by slight variations in party strengths, and government coalitions in Denmark have usually been close to minimum in size.¹³ Viewed in this light, electoral thresholds are indeed important for political stability.

In summing up at this point, we may conclude that the observations of Lipset and Rokkan are confirmed: the parties of the 1920s have by and large kept theii dominant position in the electorate, in Parliament, and, we might add, at the cabinet level par excellence. In that sense the party system has been remarkably stable. This conclusion holds true until the end of the 19605. From that time on, a modification is needed to take into account the new Socialist People's party, to which I shall return later.

To equate political stability with long-term 'aggregated' support for the four

old parties is based on an assumption about the behavioral relations among these parties. Political stability can only be inferred to the extent that the four old parties constitute a closed interaction and decision system, capable of resisting attempts at intrusion by new parties. The establishment of electoral barriers for small parties is a good illustration. In the following section I shall report on some findings from the legislative arena.

First, however, I want to refer to the main conclusions of a recruitment study of the Danish Folketing from 1849 to 1968 by Mogens N. Pedersen,¹⁴ since it relates to the topic of this paper in an interesting way. Pedersen defines four basic variables to be used in analyses of the personal circulation of a legislative elite over time: (1) permeability is the ease with which aspiring individuals can become members of the elite; (2) volatility is the extent of change in the personal composition of the elite over time; (3) continuity exists if the members of the elite tend to have a stable position in relation to the legislature and the electorate after first having acquired the position; and (4) security of tenure is the extent to which replacement of members of the elite happens owing to death, physical aging, etc. of the members. He then defines an open elite as an elite that is permeable, volatile, discontinuous, and insecure, whereas a closed elite is nonpermeable, nonvolatile, continuous, and secure. The four basic variables are then measured by a total of fifteen different indicators. Pedersen's general conclusion is that 'the Danish legislative elite has moved from relative openness to relative closedness during the 120-year period'.¹⁵ What is perhaps most interesting as regards this paper is that the changes in the indicators toward a closed legislative elite occur, roughly speaking, around 1900 and during the following two decades, i.e. at the time when the modern Danish party system was established, including the dominance of the four old parties. Consequently, it makes sense, at least to me, to view the closing of the parliamentary elite as reflecting the establishment and ensuing stabilization of the 'four old parties' system. Changes in the composition of the legislative elite are channeled through the parliamentary parties, i.e., to an overwhelming extent, the four old parties. And if the relative strengths of these parties are fairly stable from one election to the next, a constraint is put on all indicators of elite openness.

Finally, then, we can move on to the next question, which is: Do the old parties constitute a 'closed elite' in the handling of parliamentary affairs? How do they behave vis-å-vis the new parties that managed to overcome the electoral threshold erected by the old parties? We know that old parties, practically speaking, have monopolized cabinet posts, but what about normal legislative work?

3. Old and New Parties in Parliament

Let us begin with the plain but important fact that the old parties have been actors on the parliamentary scene throughout the whole period. On the other hand, new parties have come and gone 11 times since 1920 (Table III). There is considerable variation in their length of stay in Parliament, however, with the Justice party

(single tax and liberal free trade party) and the Communists leading the record with 34 and 28 years, respectively. Both parties lost their seats in 1960, when they were replaced by the right-wing Independents and the leftist Socialist People's party. The Independents soon lost ground again, as did later a new Liberal Center party (formed by dissatisfied members of the Liberal party) and the Left Socialists (a party split off from the Socialist People's party), whereas the Socialist People's party apparently has come to stay.

To find indications for the existence of a more or less closed legislative old party elite involves behavioral comparisons of old and new parties in Parliament. Unfortunately, we do not have appropriate data for the whole 50-year period. But some data on party behavior in the Folketing have been collected for the period 1953-70, which allow us to compare the old parties with five new parties¹⁶ (Justice, Communist, Socialist People's, Independent, and Liberal Center) in various respects.

The comparison has to take into account that new parties are always noncabinet parties (except for the Justice Party 1957-60, which, in effect, behaved like an old party in that period). Therefore, their behavior must be compared with that of old parties in opposition at any given time. The comparison carried out is threefold, corresponding to three important stages of the parliamentary process: political initiatives, committee work, and voting behavior in final divisions.

Legislative Initiatives

Formal rules confer several rights of legislative initiatives to members of Parliament: to propose bills and resolutions (including amendments), to move motions concerning the agenda,¹⁷ to initiate interpellation debates, and to ask questions to ministers. We therefore need to know the extent to which these formal powers are used by various members and parties and, especially, whether there are significant differences with respect to new and old parties.

Before we can test for this we must decide on a standardized way of comparing new and old opposition parties, since they vary in number, size, and periods of representation. As to size, we know that new parties generally have few members in Parliament. Nevertheless, I have counted all parties without assigning weights according to membership size in the calculations below. The rationale is that, for the purpose of studying differences between old and new parties, it is preferable to use numbers that might be interpreted as underestimating rather than overestimating differences in levels of initiation activity. The only exception made concerns the numbers of questions asked, since this type of activity is much more personal (even if the members inform their party fellows in advance about the questions they are going to ask, which is usually the case) than the other types of legislative initiatives that are considered strictly party matters. The standardization was done by dividing the actual numbers of initiatives by the number of opposition-party sessions.₁₈ Finally, I should emphasize that only 'pure' opposition party initiafives

are included, which is to say that proposals cosponsored by government parties
have been excluded. The results of calculations for six indicators are shown in
Table IV.

The table shows pronounced differences with regard to initiatives taken by new and old opposition parties. On all measures of formal initiatives new parties are more active than old parties (although the difference is vanishing with respect to interpellations), notwithstanding the fact that new parties have less resources at their disposal. The most plausible solution of this paradox is that small new parties have no other way of using their resources in legislative politics, which leads us to the topic of the role and structure of the committee system in the Danish Parliament.

Committees in the Legislative Process

The bulk of legislative work is handled by a system of specialized committees. From 1953 through 1970 the total number of bills and resolutions introduced was 3267, out of which 2800, or 85.7 %, were referred to a committee for consideration before any final decision was taken. The committee meeting is a forum for bargaining and compromise among the parties, or, at any rate, the place for ratification of agreements worked out among the major parties if informal bargaining has occurred outside the committee structure. Generally speaking, if one knows the outcome of committee discussions on a particular legislative item, one also knows the final legislative product because of the cohesive nature of the parties.¹⁹

Consequently, the party composition of committees is of great importance. In the period 1953-70, new parties, except the Socialist People's party, were not able to obtain seats in committees by their own votes. The data show that small parties were admitted to a committee in less than 1 % of the cases as an act of charity by old parties.²⁰ The committee structure thus is seriously biased against small parties, which to a large extent are excluded from participation in normal parliamentary work, no matter how many proposals they introduce or how much they speak up in floor debates.²¹ They always find themselves confronted with a fait accompli, which they may oppose or support, but not amend. That is why we

find a relatively large number of amendments moved on the floor by new parties (Table IV, 3). Small parties do not have the prior opportunity to modify bills or resolutions in the committee as do the old parties. At the same time the exclusion of small parties from committees presumably tends to decrease the motivation of their members to support proposals when the final vote is called.

Committees are also important from a nondecision-making point of view. Bills and resolutions are rarely defeated by a vote but, if not passed, simply killed by committee inaction. The only significant departure from this procedure occurred to several Communist proposals which were demonstratively voted down immediately or, at least, before a final vote was due.

Voting Behavior

Earlier legislative research has dealt nainly with the voting behavior of parliamentary parties in the final divisions on government bills that were passed.²² These studies convey a picture of the legislative party system in which any of the old parties in opposition is always closer to the government party(ies) than any of the new parties (with the single but notable exception of the Socialists in 1966-68), thus supporting the concept of the four old parties as a permanent nucleus in the legislative system. Government bills constitute by far the majority of proposals for legislative processing. Nonetheless, final divisions on government bills are only a subset of all final votes taken (2219 of 2624 in 1953-70); it is therefore of interest to know whether the above picture is upheld with the complete data base of final divisions on all types of proposals. We get the answer in Table V, which, for each cabinet period and for each opposition party, gives the percentage of divisions in which the party voted on the winning side ('yes' if proposal was passed, 'no' if defeated). The percentages in Table V show first and foremost that old

opposition parties generally vote more often on the winning side than do new opposition parties, with the only noteworthy exception being the Socialists in 1966-68, when the party was more frequently on the winning side than all of the three old opposition parties. Secondly, none of the old parties has ever been below the 80 % mark, whereas the Justice party, the Communists, and the Independents were far below that percentage.

The data on initiatives, party composition of committees, and voting behavior cannot prove the existence of a closed legislative elite of the four old parties. Viewed together, however, they tend to support the idea of a very important distinction between old and new parties. With a few modifications the distinction at the same time divides large and small parties, or, to put it another way, separates 'in' parties from 'out' parties.

There are two main difficulties with this way of summarizing. One is that the Justice party participated in a coalition government 1957-60. A study of the intricate deliberations and tactical maneuverings that preceded this quite surprising cabinet formation shows that the Justice party was included in the government, not because the Social Democrats or the Radical Liberals really were happy about it but because it turned out to be the only way in which the two parties could achieve their basic goal: to keep Liberals and Conservatives out of office.²³ As already mentioned, the Justice party soon became an 'out' party again, since it lost all seats at the following election. The other difficulty stems from the unique position that the new Socialist People's party has acquired during the 19605. It will be necessary later to devote most of a section to this development.

4. Government Stability and Performance

There are several aspects to the concept of political stability. So far we have seen that the Danish party system has been highly stable in terms of voter support for the four old parties. They have almost succeeded in retaining the overall level of support they enjoyed in the 19205. Only the (Agrarian) Liberal party support has declined. In terms of parliamentary seats the four parties have done even better owing to the effects of electoral thresholds. Although eighteen new parties have attempted to gain representation (eleven successfully for at least one election period), today only the Socialist People's party is represented in Parliament in addition to the four old parties. The net result therefore is a breakthrough of the Socialists at the expense of decreasing Liberal strength. There also are clear indications that legislative policy has basically been made by the four old parties, since new small parties are not able to overcome the 'legislative threshold' that exists, particularly for committee membership. This has obviously been another important source of political stability.

Stability can also be measured in terms of cabinet durability. J. Blondel found
that the average duration of Danish cabinets in the period 1945-65 was 2.1 years,
which placed Denmark far down at the thirteenth place among seventeen Western

democracies, with only Belgium, Finland, Italy, and France experiencing more cabinet instability.²⁴ This result is misleading, however, since it rests on a definition of government that is inappropriate in the Danish case. Blondel sees a change of government not only if a change in party support for a government occurs but also if a change of prime minister takes place, even if the cause is nonpolitical, such as ill health or death.²⁵ It is invalid to justify such an operational procedure with the argument that a change of prime minister in Italy and France normally indicates that a governmental crisis has taken place! In fact Denmark had six different governments in the period covered by Blondel's article (using only the first criterion mentioned as instrument of measurement), which gives an average durability of 3.1 instead of 2.1 years. The mean duration of all governments from 1920-71 is 3.5 years, cf. Table VI, which also shows that minority governments have a shorter life expectancy than majority governments, although the average still is about 2V2 years.

There is more to government stability than sheer durability, however. First, none of the cabinets resigned because of internal disagreements. Secondly, cabinet formations are highly predictable in so far as their parliamentary support coalitions are formed by parties adjacent to each other on the traditional left-right axis.₂₆ Finally, we need to know something about government performance in addition to duration.

Mogens Pedersen calculated 'percentages of accomplishment' (the percentage of government bills passed of all proposed) 1945-65. He found that majority governments had a high rate of accomplishment and minority governments a considerably lower one.²⁷ The values of the same indicator of government performance are listed in the first row of Table VII for the period 1953-70. These values underestimate government performance, however. First, they should be adjusted to take into account that sometimes an election is called during an ongoing Folketing session, with the consequence that all bills in the legislative machinery are dropped. The results of this adjustment are found in the second row of Table VII. Secondly, some bills do not reach the final stage in the law-making process simply because they are introduced late in session, and, usually, they are not expected to be passed right away either. Adjustments for this are made in the third row of the table by cutting out the bills that were introduced later than March 31.

The outcome of these adjustments still indicates a difference in legislative performance between majority and minority governments. Of majority government bills, 95 %-97 % are passed, whereas two of the three minority governments performed significantly less well, with only 82-84 % passed bills. The minority government of 1966-68, on the other hand, performed almost as efficiently as a majority government in terms of getting bills passed. This was the period of intimate cooperation of the Social Democratic government with the Socialist People's party.

The purpose of this paper so far has been to describe various ways in which the Danish party system, and thereby also the wider political system, has been characterized by stability over the past 50 years. Although some of the data are rather fragmentary, a set of interrelated conclusions can be drawn: (1) a high degree of stability of voter support for the four old parties and a corresponding stability of the parliamentary party system; (2) stability in legislative policy-making, which basically has been controlled by the same four old parties alone until the end of the sixties; (3) high predictability in cabinet formation; (4) fairly long average duration of governments; and (5) very high degree of legislative performance by majority governments and fairly high by minority governments.

I shall readily admit that appropriate comparative data might show that there is nothing unique about the long-term stability of Danish party politics, but that does not alter the conclusion that such a pronounced stability does exist. The two remaining questions then are: How is past stability to be explained, and Why did major changes occur toward the end of the 19605?

5. Causes of Past Stability and Recent Change

The main problem is to explain why new parties generally fail to keep or increase their strength once they have obtained representation in Parliament, or, viewed from the other angle, how to explain that we find a continuation of a party system established early in the century when PR actually should give new parties a fair chance to capture votes from the old parties. It is true that the four old parties from the outset constituted an inner circle of the legislative system, which, structurally, is severely biased against new and small parties. But this does not fully explain why new parties typically remained small parties even if they had managed to pass the electoral thresholds.

In 1956, Dankwart Rustow concluded that:²⁸ 'The persistence of the four-party pattern throughout Scandinavia is to be attributed to the fact that it closely reflects the social structure in each country and that social and economic differences today provide the main stimulus for party division. The Conservatives, Agrarians (i.e. Agrarian-Liberals in Denmark), and Socialists represent the interests of employers, farmers, and workers, respectively. Those middle class elements not readily identified with any of these groups have provided the chief support for the Liberal (i.e. Radical Liberal in Denmark) parties.' Lipset and Rokkan ask questions similar to those above in their comparative analysis. In their attempt to find an answer they suggest that one has to start out from an analysis of 'old' and 'new' parties:²⁹ 'the early mass parties formed during the final phase of suffrage extension, and the later attempts to launch new parties during the first decades of universal suffrage. It is difficult to see any significant exceptions to the rule that the parties which were able to establish mass organizations and entrench themselves in local government structures before the final drive toward maximal mobilization have proved most viable. The narrowing of the "support market" brought about through the growth of mass parties during this final thrust toward full-suffrage democracy clearly left very few openings for new movements.' Considering these two quotations together, we end up with the following explanation of the questions posed: At the stage of political development where our analysis starts, the four old parties had already attained stable relationships with the most important groups of voters, such as farmers, workers, and employers, thus virtually emptying the 'support market' and leaving no major 'openings' to new parties.

This hypothesis, which also implicitly explains the decline of the Liberal party, certainly makes much sense, but before applauding we must deal with some possible difficulties. Throughout this paper I have talked about new/small parties on the one hand and old/big parties on the other. In fact, three of the new parties were not really that small at points in time: the Communists around 1945, the Justice party in the early fifties, and the Socialist People's party since the 1960s (cf. Tables I and III). Obviously, then, there must have been some opening in the electorate for these parties. The surge of the Communist vote in 1945 is usually explained by the popularity the party enjoyed because of the job it had performed within the Danish liberation movement, which, however, disappeared with the

advent of the cold war. The success of the Justice party in 1950 is often attributed to a 'protest party' effect by which the party temporarily benefited from disappointed non-Communist workers as well as from liberal and conservative voters.³⁰ Unfortunately, there is a research problem with respect to these two parties in that no adequate data are available to test hypotheses of this sort. This leaves us with the Socialist People's party and the changes that occurred in the party system during the 19605.

Retrospectively, the election of 1960 looks almost like a landmark. First, it resulted in a bare majority for the Social Democrats and Radical Liberals, the two old alliance partners, who formed a coalition government for the last time (as far as one can tell today). Secondly, the two oldest of the new parties, the Communists and the Justice party, disappeared completely from the Folketing. Thirdly, the Socialist People's party, launched by the former Communist leader, as the first new party in Danish history, obtained 6.1 % of the vote and 11 seats in the Folketing at their first try. The party suffered a minor setback in 1964 but more than regained its support in 1966, with 10.9 % of the vote and 20 mandates. This result led to a major change in Social Democratic party strategy. The Social Democrats had applied a strategy of encapsulation against the Socialists, a tactic that had earlier worked against the Communists, but now they switched to a strategy of cooperation because jointly the two parties could command a majority in Parliament, which had not been the case previously. The Radical Liberals, in turn, were persuaded by this development to approach Liberals and Conservatives, and from 1968 through 1971 these three old parties formed a coalition government. Since 1971 the Social Democrats have again been in office, supported by the Socialist People's party. In other words, the classical four-party system is now, in effect, replaced by a five-party system with new coalition patterns because of the emergence of the Socialist party. Foreign policy is handled more or less in concert by the four old parties, whereas domestic issues often align Social Democrats and Socialists against the Radicals, Liberals, and Conservatives. Viewed as a new party, the Socialist party has shown a remarkable viability, and it has moved close to the decision-making center through cooperation with the Social Democrats. It would indeed be interesting to know how the party managed to accomplish what the previous history of the party system made believe was impossible. It would be unwise to pretend that there is a straightforward answer, but some important points can be made.

The way in which the Socialist People's party presented itself to the voters can be described, perhaps with some oversimplification, as follows: In domestic policy areas the party offered a program appealing to the working class, particularly to those workers who did not feel attracted to orthodox communism but at the same time considered the actual policies of the Social Democrats, made by compromises with the 'bourgeois' parties, to be too far away from their original socialist principles. In the foreign and defense policy domain the party presented a strongly articulated rejection of Danish membership of NATO and the EEC, probably the two most salient foreign policy issues of the 19605.

As will be recalled, defense had been a major issue among the old parties in the interwar period. During the 19305, however, the Social Democrats began changing their attitude toward Danish defense owing to the developments in Nazi Germany. After the Second World War the Social Democrats were prepared to join NATO, whereas the Radical Liberals still opted for Danish neutrality and a limited defense budget. The Radicals therefore voted against ratification of the North Atlantic Treaty. During the fifties, however, the Radical party also retreated from its original foreign policy philosophy. In sum, if a question of 'yes or no to NATO' was asked to the four old parties in Parliament in 1960, they would all answer yes, although some parties would speak more loudly than others. The same alignment occurred on the EEC issue.

In 1960, the Socialist People's party thus stood as a democratic, socialist alternative to Communists and Social Democrats. In foreign and defense policy the party could take over the space left open by the Radical Liberal change of policy. We cannot prove that these factors explain how the party found an 'opening' to the 'support market' because data are lacking for the 19605. But we do have data from the 1971 national Danish survey that tell us something about relevant characteristics and attitudes of socialist voters compared with voters who preferred one of the four old parties.

First, compared with the old parties, we would expect the Socialist People's party to draw proportionally more support from young voters than from older voters, and more from workers than from any other occupational category. Tables VIII and IX provide the data to test these hypotheses.

Concerning age, the hypothesis is clearly confirmed. The age distribution of socialist voters shows that support for the Socialist party decreases with increasing age. Further, one third of socialist voters are within the age group of 21 to 30 years, and 30 % within the 31- to 40-year category. This actually means that, in 1971, no less than 63 % of the Socialist party's support came from voters who voted in an election for the first time in 1957 or later.

As to occupation, Table IX shows the expected result: the Socialist People's
party, on a smaller scale, is just as much a workers' party as is the Social Democratic

party; about 53 % of its voters are workers, compared with 51 % of the
Social Democratic voters.

Turning next to the questions of ideology and foreign policy attitudes of socialist voters, we can first refer to a. study by Ole Borre³¹ on party and ideology, which has the same 1971 survey as data source. He cross-tabulated party choice by ideological beliefs and found that Socialist People's party supporters on the average were the most radical, or leftist, in a set of eight questions designed to measure ideological orientation on a domestic issue left-right scale. As we would expect, too, the average location of the four old parties' voters followed suit in the traditional order.

As to the two specific foreign policy issues mentioned - Danish membership of
NATO and of the EEC - the 1971 data also provide an opportunity to determine
differences among socialist voters and voters who chose one of the four old parties.

Table X indicates a very firm rejection of Danish membership in NATO by
Socialist People's party voters. They are about as overwhelmingly against membership

as Liberals and Conservatives are for, at the other end of the spectrum.
In between we find the Social Democrats and the Radicals, with the Social Democrats
as the least enthusiastic pro-NATO party of the two.

Data on the EEC membership issue are reported in Table XI. As with the

NATO issue, the socialist voter differs dramatically from other voters in their opposition to the EEC. Otherwise the picture is very much similar to the earlier one, except that Social Democrats are even less enthusiastic about the EEC and that the Liberals this time are more positive than the Conservatives, presumably reflecting the economic benefits to agriculture from EEC membership. The socialist voters also reported more interest in the question of Denmark's entry into the EEC than other voters, 53 % answering 'very much' interested, compared with 31 % to 44 % for supporters of the four old parties.

Because of the small sample of socialist voters, we ought to be cautious in drawing conclusions about the characteristics of the total population of Socialist People's party voters. On the other hand, the data indicate that socioeconomic variables cannot explain why some voters choose the Socialist party rather than the Social Democrats. Ideology seems to be the decisive factor in this respect.

Nevertheless, in Table IX above we still find clear relationships between social groups, as measured by occupation, and party choice along the traditional lines with respect to the four old parties. We can supplement that table by turning it around and asking about the distribution of party choice within categories of occupations. This has been done in Table XII, where housewives and 'other positions' have been omitted, as have all 'not ascertained' on party vote, which reduces the sample size to 860. We find that farmers vote heavily Liberal, that workers strongly prefer the Social Democrats, and that self-employed in urban industries prefer the Conservatives; the Radical Liberals is not the largest party within any of the categories, but it obtains between 10 % and 20 % within all groups. All of this is quite in line with the hypothesis proposed to explain the longterm stability of the party system. That much said, however, we must also realize

that there is by no means a one-to-one correspondence between occupational position and party choice. And further, whereas both farmers and workers are declining in numbers, white-collar today has become a very important voting group that is not as strongly attached to a particular party as other groups.

In sum, then, although the hypothesis seems to explain long-term stability better than any other, we should neither assume that everything is explained nor that stability will continue. The party system of the 19605, on the one hand, reflects the cleavage structure of the 19205; on the other hand, the changes that occurred in the 1960s might portend further structural developments in the party system of tomorrow.

6. The Election of 1973: Short-Term Protest or Permanent Change?

The preceding analysis was based on data covering half a century of Danish party politics from the 1920s onwards. The findings corroborated the hypothesis of Lipset and Rokkan that the party system of the 1960s basically reflects the cleavage structure of the 19205. The persistent support for the four old parties is at the same time the key to understanding the remarkable stability of the political system throughout the five decades.

Around 1970, however, the party system had changed in important respects: the Socialist People's party had been added, and, partly as a result of that, the patterns of party coalitions were different from the traditional ones. Further, the electorate of the 1960s showed clear signs of increasing mobility and volatility, not only by returning the new Socialist party but also by more than doubling the Radical Liberal vote from the November 1966 election to the January 1968 election and by turning out in higher numbers than ever before (86 % to 89 %). Finally, there was a steady increase in voter support for parties that did not pass the electoral threshold of 2 % (cf. Fig. 1).

On the basis of these findings, therefore, one might reasonably interpret the situation of the early 1970s as one of continuity and change, with continuity as the dominating component. Before the end of 1973, however, the party system was changed almost beyond recognition.

In the spring of 1973 a new 'progressive party' suddenly made its appearance. According to public opinion polls it received support from no less than 25 % of the voters at one time. The party, which initially was a one-man show by a lawyer named Mogens Glistrup, used slogans that appealed to people who were getting tired of increasing tax burdens, red tape bureaucracy, alleged misuse of social welfare money, and the like. At the same time, according to the polls, the four old parties were losing support, especially the governing Social Democratic party.

On the basis of rational analyses, therefore, it would be unwise to call an election. But then another unforeseen event occurred. The government, supported by the Socialist People's party, had only a bare majority in Parliament. Thus, when a 'right-wing' member of the Social Democrats left the party as a protest against the 'leftist' policies of Social Democrat and Socialist People's parties, particularly in the area of property taxes on house ownership, a stalemate government-opposition situation arose. The government decided that this situation was untenable and called an election, with agonizing results, as shown in Table XIII.

The new Folketing is indeed very different from that of 1971. For the first time all established parties suffered severe losses simultaneously. The four old parties, which had commanded 90 % of the seats before the election, were returned with only 60 % of the seats. Instead of five there are now ten parties in the Folketing. Two of the small new parties are well known from the earlier history of the party system, viz. the Communists and the Justice party. The third small new party is the Christian People's party, which by only a few votes failed to obtain seats in the 1971 election. Neither the fourth nor fifth new party is a small party, however. The Center Democrats, led by the Social Democratic secessionist mentioned

above, captured 14 seats, and the Progressive party obtained 28 mandates in the
Folketing.

Within our longitudinal perspective, the significance of the election is not that it produced a ten-party Folketing (after all, eight or nine parties have been represented before) but that all five established parties suffered setbacks and that two parties of considerable size were added (Center Democrats, Progressives). Tn other words, the Folketing is now more fractionalized than ever, with consequences for the possibilities of majority-building. The best way to visualize this is to graph the fractionalization³² of parliament throughout the period covered, cf. Fig. 2.

At the moment of writing (January 1974), survey data on the election are not available, and, although a lot of factors have been suggested to explain the surprising results, I shall refrain from attempts at interpreting the motives of voters who switched to new party alternatives. Suffice it to say that the election had all the characteristics of a massive protest by a huge number of people, who, since the rapid increase in the standard of living during the 19605, had taken the services of a modern welfare state for granted but had not yet quite realized its costs. It is not yet possible to tell whether this protest represents a short-term phenomenon or a more basic realignment within the electorate, but it is not unlikely that there will be some sort of 'reinstating' election the next time, perhaps even before this paper is published.

---

NOTES

1. The classical study along these lines is of course Maurice Duverger's book on political parties. See Maurice Duverger. Party Systems Book II of Political Parties. London, Methuen & Co., 2nd ed., 1959.

2. Some authors, like Gabriel A. Almond, preferred to put the labels 'Anglo-American' and 'Continental European' political systems on those systems that were characterized by the presence or absence of a homogeneous, secular political culture and a highly differentiated role structure. As Arend Lijphart points out, however, Almond's typology hardly differs from classification by the number of parties in terms of democratic stability; cf. Arend Lijphart 'Typologies of Democratic Systems', pp. 46-80, at pp. 55 f. in A. Lijphart (ed.) Politics in Europe. Englewood Cliffs, N.J., Prentice-Hall, 1969.

---

2. Some authors, like Gabriel A. Almond, preferred to put the labels 'Anglo-American' and 'Continental European' political systems on those systems that were characterized by the presence or absence of a homogeneous, secular political culture and a highly differentiated role structure. As Arend Lijphart points out, however, Almond's typology hardly differs from classification by the number of parties in terms of democratic stability; cf. Arend Lijphart 'Typologies of Democratic Systems', pp. 46-80, at pp. 55 f. in A. Lijphart (ed.) Politics in Europe. Englewood Cliffs, N.J., Prentice-Hall, 1969.

3. In Sigmund Neumann (ed.) Modern Political Parties. University of Chicago Press, 1956, p. 191.

4. Giovanni Sartori. 'European Political Parties: The Case of Polarized Pluralism', pp. 137-176 in Joseph LaPalombara & Myron Weiner (eds.) Political Parties and Political Development. Princeton, N.J., Princeton University Press, 1966. See also G. Sartori. 'The Typology of Party Systems - Proposals for Improvement', pp. 322-352 in S. Rokkan & E. Allardt (eds.). Mass Politics: Studies in Political Sociology. New York, The Free Press.

5. Article mentioned in note 2 above.

6. This conclusion is also implicit in empirical works like the following: J. Blondel. 'Party Systems and Patterns of Government in Western Democracies', Canadian Journal of Political Science 1968: 180-203. Hans Daalder. 'Cabinets and Party Systems in Ten Smaller European Democracies', paper to IPS A Round Table, Torino, September, 1969. Michael Taylor and V. H. Herman. 'Party Systems and Government Stability', American Political Science Review LXV (1), 1971: 28-37.

7. Seymour M. Lipset and Stein Rokkan. 'Cleavage Structures, Party Systems, and Voter Alignments', p. 50 in Lipset and Rokkan (eds.) Party Systems and Voter Alignments. New York, The Free Press, 1967, pp. 1-64.

8. Ibid., p. 14.

9. But not on all. For example, the Liberals and the Radical Liberals were strongly in favor of free international trade (to the advantage of agriculture), whereas both the Social Democrats and the Conservatives were more inclined to use protectionist trade policies (to the benefit of domestic industry and employment).

10. The positions of the four old parties on the defense policy dimension could be measured meaningfully on a ratio scale these days. In 1922 a defense committee reportedon a reorganization of the Danish defense system; each of the parties presented a proposal. Below are the costs of implementing the four proposals (in millions of D. kr.): Soc.Dem. Rad.Lib. Lib. Cons. 7.5 22 ,40 49 Since Liberals and Conservatives commanded a majority in Parliament, the outcome was a compromise between the two parties — at the cost of 44 million D.kr., of course!

11. The Danish election system is perhaps the most complicated in the world. Basically, it assures that the parties get the share of seats that corresponds to their share of the vote by means of the method of the largest remainder. The actual allocation of mandates occurs in two waves: first 'constituency mandates' are given to the parties according to the modified Ste-Lagues method (before 1953, by the d'Hondt method). Secondly, supplementary mandates' are allocated to ensure that the first allocation does not deviate from proportionality. To get a share of the supplementary mandates, however, a party has to fulfill certain requirements that have changed over time: from 1920 through 1953 a party should either have obtained at least one constituency seat (which is always difficult for small parties unless they have local strongholds) or, within one of the three main regions (Copenhagen, Jutland, and the Islands), have obtained at least as many votes as on the average were cast per mandate nationally. From 1953 through 1961 the second qualification was applied to all three regions instead of only one, which of course hurt regionally unbalanced new parties. At the same time a third rule was introduced, by which parties could also get supplementary mandates if they had obtained 60,000 votes in the country as a whole. Finally, from 1961 the second qualification rule was loosened a bit, so that it applied to only two of the three main areas, and the third rule was changed to 2 % of the total vote. In addition to these thresholds, new parties not represented in Parliament have to collect a certain number of signatures to qualify for participation in elections (10,000 until 1965, thereafter corresponding to the number of votes cast at the latest election per mandate, which is substantially more than 10,000).

---

11. The Danish election system is perhaps the most complicated in the world. Basically, it assures that the parties get the share of seats that corresponds to their share of the vote by means of the method of the largest remainder. The actual allocation of mandates occurs in two waves: first 'constituency mandates' are given to the parties according to the modified Ste-Lagues method (before 1953, by the d'Hondt method). Secondly, supplementary mandates' are allocated to ensure that the first allocation does not deviate from proportionality. To get a share of the supplementary mandates, however, a party has to fulfill certain requirements that have changed over time: from 1920 through 1953 a party should either have obtained at least one constituency seat (which is always difficult for small parties unless they have local strongholds) or, within one of the three main regions (Copenhagen, Jutland, and the Islands), have obtained at least as many votes as on the average were cast per mandate nationally. From 1953 through 1961 the second qualification was applied to all three regions instead of only one, which of course hurt regionally unbalanced new parties. At the same time a third rule was introduced, by which parties could also get supplementary mandates if they had obtained 60,000 votes in the country as a whole. Finally, from 1961 the second qualification rule was loosened a bit, so that it applied to only two of the three main areas, and the third rule was changed to 2 % of the total vote. In addition to these thresholds, new parties not represented in Parliament have to collect a certain number of signatures to qualify for participation in elections (10,000 until 1965, thereafter corresponding to the number of votes cast at the latest election per mandate, which is substantially more than 10,000).

12. In fact, only the Liberals got an advantage, which was due to the then applied d'Hondt method of allocation of constituency mandates, combined with the rule that parties that in this first allocation had obtained more seats than they should have by the proportional standard could keep these surplus seats.

13. Detailed information is provided in Erik Damgaard. 'The Parliamentary Basis of Danish Governments: The Patterns of Coalition Formation', pp. 30-57, at p. 37 in Scandinavian Political Studies, vol. 4. Oslo, Universitetsforlaget, 1969.

14. Mogens N. Pedersen. 'The Personal Circulation of a Legislature: The Danish Folketing 1849-1968'. Mimeograph, Institute of Political Science, University of Aarhus, 1972.

15. Ibid., p. 18.

16. The sixth party, the Left Socialists, had to be omitted because of a missing data problem. The party's four-man group soon split into two groups, for which complete data are lacking. They are included in the numbers on formal initiatives, however.

17. An agenda motion is a particular type of motion that is used by the Folketing as a means to articulate opinions or demands, especially to the government, with a motion of censure as its extreme form.

18. The number of party-sessions for old opposition parties is 39, compared with 32 for new parties.

19. This is shown in my paper, 'Party Coalitions in Danish Law-Making 1953-1970', European Journal of Political Research 1, 1973: 35-66, especially pp. 39-42.

20. Almost all situations occurred in 1953-57 and involved the Justice party, which, in about half of the cases, had proposed the bill or resolution itself. The party also had members on committees during its cabinet membership period 1957-60.

21. Mogens Pedersen gave a good example to illustrate the small parties' point of view. The leader of the Independents once said: 'What we want is to participate in the solution of the problems of society, and not to be kept outside the wall set up by the four old parties around themselves and the work of the Folketing'. Mogens N. Pedersen. 'Consensus and Conflict in the Danish Folketing 1945—65', p. 165 in Scandinavian Political Studies, vol. 2. Oslo, Universitetsforlaget, 1967.

22. Mogens N. Pedersen, article quoted in note 21. Mogens N. Pedersen, Erik Damgaard, and P. Nannestad Olsen. 'Party Distances in the Danish Folketing 1945-68', pp. 87-106 in Scandinavian Political Studies, vol.. 6, Oslo, Universitetsforlaget, 1971. Erik Damgaard. 'Party Coalitions in Danish Law-Making', references in note 19.

23. Tage Kaarsted. Regeringskrisen 1957. Universitetsforlaget i Aarhus, 1964.

24. Jean Blondel. 'Party Systems and Patterns of Government in Western Democracies', Canadian Journal of Political Science 1, 1968: 191.

25. Ibid., p. 190.

26. Erik Damgaard. 'The Parliamentary Basis of Danish Governments', op. cit., and 'Party Coalitions in Danish Law-Making 1953—1970', op. cit.

27. Mogens N. Pedersen. 'Consensus and Conflict in the Danish Folketing 1945—65', op. cit., p. 151.

28. Dankwart A. in. 176

29. Lipset and Rokkan, op. cit., p. 51.

30. E.g. Aksel Larsen. Vælgere på vandring. Universitetsforlaget i Aarhus, 1967, pp. 28-44.

31. Ole Borre. 'Party and Ideology in Denmark'. Mimeograph, Institute of Political Science, University of Aarhus, 1973.

32. The concept of fractionalization is that of Douglas W. Rae (The Political Consequences of Electoral Laws. New Haven, Yale University Press, 1967, pp. 53-58, 62). It has the property of combining the number of parties and their size in a single continuous measure that varies between zero (no fractionalization) and 1 (approaching total fracticnalization). It is defined by '—(A>!) where s is the party's share of the seats.