In theorizing about political parties, one major approach has been to conceive of parties as structures or agencies fulfilling specific functions in the political system. Among the functions attributed to parties are structuring and mobilizing the vote, political education and activation, representation and/or aggregation of interests, leadership selection, conducting the government, and policy formulation. Functional hypotheses of party decline visualize party crisis and the process of party decomposition as the failure or lessening capacity to perform these functions. These hypotheses of party decline often highlight two problems: 1) changing instrumentalities which affect how parties can carry out their functions and 2) the emergence of rival structures or forms of political organization which encroach upon and even usurp the functions of parties.

Resumé

Functional hypotheses of party decline define parties as fulfilling specific functions in the political system and relate party decay to the inability or diminished capacity of parties to perform these functions. This article examines two major hypotheses of this sort by assessing their relevance in the case of the Danish, Norwegian and Swedish Social Democratic parties. First, it deals with the thesis of the declining functionality of the mass membership party. Second, it discusses the rival structure hypothesis which argues that rival forms of political organization pose a threat to parties by encroaching upon their functions. These arguments are confronted with available empirical data. In several instances, the empirical evidence points to inadequacies in the hypotheses, or directly confutes them. These weaknesses of the hypotheses are discussed, and alternative interpretations are offered.

---

* This article is an expanded version of parts of a paper presented at the ECPR Workshop on Problems of Party Government, Aarhus, Denmark, 1982. I would like to express my thanks to Gunnar Sjdblom, University of Copenhagen, Olof Ruin, University of Stockholm, and Barbara Haskel, McGill University, for their helpful criticisms and comments. For a discussion of structural and strategic hypotheses of party decline and the Scandinavian Social Democratic parties, see Sainsbury 1982a and 1982b.

This article focuses on hypotheses related to these two problems. It examines the arguments concerning the declining ability of parties to fulfil functions — or what I shall call declining functionality — because of changing instrumentalities and the emergence of rival structures. These arguments are examined by assessing their applicability in the case of the Social Democratic parties in three of the Scandinavian countries — Denmark, Norway and Sweden. Indirectly the discussion also attempts to determine how much light these hypotheses shed on the difficulties experienced by these parties during the 19705.

Declining Functionality and Changing Instrumentalities

One of the most influential treatments of political parties, adopting an approach which has the functions of parties as its focal point, is Leon D. Epstein's Political Parties in Western Democracies (1967). Admittedly, Epstein presents his theory as one of party development and not as a theory of party crisis. In retrospect, however, it looks increasingly like a theory of party decay with special implications for mass parties and in particular working-class parties.

He argues that mass membership during the post-war period has become less of a necessity in fulfilling the party function of vote-getting. Changing instrumentalities — new techniques in political communication, especially mass media, and new sources of financing, such as government subsidies and fund-raising campaigns — provide alternatives to mass organization. Accordingly, changing instrumentalities, and especially if successfully utilized by rival parties, will be accompanied by a decline in membership. In marshalling evidence for his thesis, Epstein points to the membership trends in the 1950s and early 1960s of working-class parties which typify the mass membership party. Membership of most socialist parties peaked in the late 1940s and early 19505; subsequently, a decline set in. He concludes that the downward trends in membership are related to the ascendancy of new campaigning techniques and the legislation of party subsidies, and that these tend to displace existing membership functions.

An additional feature of Epstein's argument is a polemic addressed to Maurice Duverger, who maintained the functional superiority of the mass party in relation to the cadre party in modern politics. Duverger speculated that the future development of parties would be the universal adoption of the former organizational form. Epstein, on the contrary, holds that the new techniques are more effective in vote-getting than mass organization, and that they represent the future. In some instances, mass membership may be dysfunctional through its resistance to the new techniques, or its very existence may prevent or delay the party in exploiting these more effective techniques.

US parties characterized by a purely electoral organization, as distinct from a mass membership organization, and heavy reliance on public relations campaigning and other sources of financing than membership dues are held up as a model of probable party development.

Epstein's view of the effectiveness of new techniques causes him to underestimate the problematic features of a weakening in the organizational strength of parties and its possible consequences for the party function of vote-getting. The diminished capacity of US parties to structure and mobilize the vote in the 1970s casts serious doubts upon several of Epstein's assumptions and raises the question of whether his theory should not be viewed as one of party decay.

Let us look more closely at his argument concerning the decline in mass membership of socialist parties to see to what extent his thesis is substantiated by the experiences of the Scandinavian Social Democrats. Is the mass membership organization of the Social Democratic parties in the process of decomposition? And, if so, to what degree is decomposition related to changing instrumentalities?

Taking the late 1940s as a baseline, both the Danish and Norwegian Social Democrats have suffered a drop in party membership. However, the decline is far more severe for the Danish party (SD). At its height (1948), the membership of the SD numbered slightly over 300,000 which represented 38 per cent of the party vote. Since then, membership has steadily fallen, and in 1977 it totalled 123,000 — or 11 per cent of the party vote (Thomas 1977, 270; Pedersen 1981, 69). The membership of the Norwegian party (DNA) peaked at roughly 200,000 at the end of the 19405. Enrollment dropped during the next decade, but the overall trend from the late 1950s to the late 1970s is stagnation.¹ In 1957 party membership was approximately 154,000 and constituted 18 per cent of the party vote. In 1977 the figures were very similar (Elvander 1980, 162f; Svåsand 1981, 16-19).

As distinct from the SD and DNA whose membership enrollment culminated in the late 19405, the Swedish Social Democratic party (SAP) has grown during the entire post-war period. In fact, its membership has almost doubled since 1945. Then party enrollment was a little over half a million and accounted for 39 per cent of the party vote; in 1974 it exceeded one million members and represented 42 per cent of the party vote (Scase 1977, 337f). Almost as impressive as the overall increase is the remarkably steady nature of the growth in party membership, although growth was more sluggish in the 19505.

Three divergent membership patterns thus emerge: sharp continuous decline (SD), decline followed by stagnation (DNA) and growth (SAP). Is there a corresponding dissimilarity in the alternatives made available to the parties through changing instrumentalities, and particularly in the cases of the Danish and Swedish parties? In both countries television began to play

a role in electioneering during the 1950s (Pedersen 1981, 72; Kronvall 1975, 640 the early 19605, for example, nearly half of the Swedish electorate reported that television was the most important source of information about election campaigns. Furthermore, both countries apply a principle of 'equal-time access', although it is applied more liberally with regard to small parties in Denmark. Thus television has provided the Social Democrats with a new, inexpensive campaign resource in both Denmark and Sweden. But the addition of television as a campaign resource and increasing reliance on public relations electioneering fail to provide an adequate explanation. In this respect, the media situation is quite similar in Denmark and Sweden, but membership trends for the Social Democrats are drastically different. And with regard to party subsidies, the outcome is exactly the opposite to the one predicted: membership growth is sustained in a country with an extremely generous system of subsidies, and a decline occurs in a country without subsidies.

Obviously many factors have contributed to the dissimilar membership trends of the SD and SAP. Here I should like to explore one possible explanation — dissimilar options or dissimilar perceptions of options concerning election strategies. However, before doing so, it is necessary to comment on a factor which is frequently cited as accounting for the differences between the two parties in their ability to recruit new members. The SD is totally reliant on individual membership, whereas members of the SAP can be either collectively or individually affiliated. Moreover, the bulk of the SAP's members are collectively affiliated (73 per cent in 1974). The importance of these different forms of membership should not be exaggerated, and two points are relevant here. First, the DNA has a similar system of collective affiliation but this has not led to increasing party membership. Second, the number of individually affiliated members of the SAP has also doubled since World War II (Elvander 165f)- Thus the form of party membership does not provide the entire answer.

How can the exceptional membership growth of the SAP be explained? Part of the answer, 1 would argue, is that patterns of electoral mobilization offered, or were perceived as offering, different strategic options to the parties, and in the case of the SAP one of the major options available put a premium on party organization.

Comparing the electoral mobilization of the three countries based on the average turnout per decade for the period 1920-80 (Table 1), we find that through the 1960s Denmark consistently ranked first. Already in the 1930s the Danish average turnout was 80.2 per cent with the Norwegian turnout nearly as high, whereas Sweden trailed behind with only 71.1 per cent. In the 1940s Denmark outdistanced both Norway and Sweden. The gap between the turnout averages declined in the 19505, and Sweden inched ahead of

Norway. The difference between Denmark and Sweden virtually disappeared
in the 19605, and the rank order changed in the 1970s with Sweden having
the highest average turnout for the decade.

Electoral participation in all three countries increased during the period (for Denmark 78 per cent in the 1920s to 88 per cent in the 19705, for Norway 68 per cent to 82 per cent, and for Sweden 58 per cent to 90 per cent). What stands out, however, are the differences in the Swedish case. First, the size of the increase in electoral participation is much greater — more than 30 percentage points compared to 10 and 14 percentage points for Denmark and Norway respectively. During the period of sixty years, Sweden moved from being the lägger to the leader. Second, Swedish electoral mobilization, as measured by average turnout per decade, exhibits a steady increase. In the Danish case there was a fall off in electoral participation during one decade — the 1950s — and in the Norwegian case participation stagnated at the 1930s level during the next two decades and it declined slightly again in the 19705.

The pattern of Swedish electoral mobilization merits closer inspection (Figure 1). A characteristic feature was spurt elections, or as they have been more aptly termed — mobilizing elections, when the turnout rate increased by several percentage points. Generally, in the next election or elections voting participation declined, but it did not drop to the level prior to the mobilizing election (Korpi 1981, Chapter 2). Mobilizing elections occurred in 1928 (+l4 per cent), 1936 (+6 per cent), 1948 (+ll per cent), 1960 (+9 per cent) and 1968 ( + 5 per cent).

With the exception of the 1928 election, the mobilizing elections were crucial to the Social Democrats' long period in government. The 1936 election confirmed and strengthened the party's position in the electorate and in office. The subsequent mobilizing elections all share a common feature: prior to the election the prospects of losing power loomed large. In this situation the party adopted a mobilization strategy to get all party sympathizers to the polls.

The lower turnout rates in Sweden, especially compared to Denmark, held
out the possibility of a successful mobilization strategy and probably contributed

to the perception of this strategic option. For example, prior to the 1948 election the voting turnout was only 70 per cent, which left nearly onethird of the electorate to be mobilized. Moreover, class differentials in electoral participation meant an even greater potential payoff for the SAP. The simple fact that the 1948 strategy succeeded, and the Social Democrats retained control of the executive, despite a grave challenge, is also significant. Such a strategy was assigned a place in the party's repertoire of successful election strategies. Furthermore, and extremely important, was the role of the mass membership organization in the implementation of the strategy.²

How does this pattern of electoral mobilization compare with the Danish and Norwegian cases? Electoral mobilization in Denmark was relatively high for the whole period, and increases in voting participation were generally very modest — at most only 2 or 3 per cent higher than the turnout of the preceding election (Figure 1). In fact, there is only one election during the entire period which might qualify as a mobilizing election. In the 1943 election the voting turnout climbed by 10 percentage points to 89.5 per cent — an alltime record. However, this impressive mobilization of the electorate was a manifestation of national unity during occupation — and not the product of intensified party competition for the control of government as in the Swedish case.

A comparison of Norway and Sweden discloses a number of similarities through the 19405. But then the parallel ends. During the years 1920-50 three mobilizing elections occurred in each country. The 1930 mobilizing election in Norway also resembled the 1928 mobilizing in Sweden inasmuch as the outcome entailed a setback for the Social Democrats. However, in the mobilizing elections of 1936 the Social Democrats were victorious in both countries. There also appear to be similarities between the 1948 Swedish election and the 1949 Norwegian election. At any rate, both elections were hotly contested (Torgersen 1962, 1600, and the Social Democrats waged successful campaigns.

During the next two decades, differences rather than resemblances stand out. First and foremost, the next Norwegian mobilizing election in 1965, contrary to the Swedish pattern, resulted in the Social Democrats' losing office. Second, during the 1950s Norwegian electoral participation declined but the DNA increased its share of the vote in both the 1953 and 1957 elections — and the party won parliamentary majorities. The voting turnout in Sweden also declined in the 19505. But as distinct from the DNA, the SAP was precariously close to losing power during the entire decade, and from 1956 onwards there was a non-socialist majority in the electorate. With the 1960 mobilizing election the SAP bettered its standing with the voters, and a socialist majority in the electorate was reinstated. When new party losses set in, the SAP mounted a mobilization campaign in 1968 which bore several

resemblances to the 1948 strategy.

In the 1970s Swedish electoral participation approached its ceiling. Mobilizing elections with major increases in voter turnout virtually became an impossibility. As a result of this change, the question has been posed whether the heyday of the mobilizing strategy has not come to an end in Swedish politics. The preconditions for a successful mobilizing strategy have been significantly altered since the 19505. The high turnout rates have meant that the number of votes which a party could hope to win by getting voters to go to the polls has diminished. At the same time, volatility in the electorate has increased during the past two decades, and the number of voters prone to switch parties has grown. Now potential gains seem greater from a strategy attempting to persuade voters to change parties than from a mobilizing strategy and, accordingly, the former type of strategy ought to take precedence over mobilization efforts (Holmberg 1981, 33).

This line of reasoning, however, overlooks several factors which continue to make a mobilizing strategy important, and especially to the SAP. Initially it should be noted that the total percentage of party switchers inflates these voters' strategic importance, at least from the vantage of the Social Democrats. Although the percentage of voters changing parties in the late 1970s approached 20 per cent (compared to 7 per cent in the mid-19505), voters making the crucial switch from a non-socialist to a socialist party or vice-versa amounted to only 5 per cent in 1979 (Holmberg). In other words, the group of voters who through changing parties might decide the outcome of the election in favour of either the socialists or non-socialists remained quite small.

Among the factors contributing to the continued importance of mobilizing efforts is a stronger proclivity among working-class voters, low income groups and the disadvantaged, compared to other groups, to stay away from the polls (e.g. Petersson 1977, 252). This tendency alone makes the mobilization of voters more essential to the Social Democrats than to the non-socialist parties. An additional factor is the keen competition between the socialist and non-socialist parties, which was especially intense during the 19705. The 1973 and 1970 elections were extremely close contests. The difference between the two blocs of parties represented in parliament was an infinitesimal .06 per cent in 1973 and under .2 per cent in 1979 — or around 3,800 and 8,500 votes respectively out of an electorate of approximately six million. In elections as close as the 1973 and 1979 campaigns, the potential payoff of getting out the vote remains very high, indeed. Finally, survey data from the 1979 election indicate that SAP sympathizers were over-represented among demobilized voters (those who voted in 1976 but abstained from voting in 1979) and other non-_voters (Holmberg 1981, 33-5). Under these circumstances, it is hardly surprising that mobilization of the electorate was a major feature of the

SAP's campaign in the 1982 election.

As long as party competition is intense, the margin of victory in elections is narrow, and Social Democratic sympathizers are overrepresented among the non-voters and hesitant voters, a mobilization strategy is likely to remain an integral part of the strategic repertoire of the SAP. The outcome of the 1982 election justifies a similar conclusion.

In sum, the Swedish experience deviates sharply from Epstein's predictions of 'party development'. Despite the growing role of television and public relations techniques in election compaigns and generous subsidies to parties, the mass membership of the SAP has not declined. Part of the explanation for the deviation seems to be found in the strategic situation of the SAP, the party's perception of its options, and its responses. In a situation of intense party competition and impending loss of power, a party is likely to utilize all available resources — both traditional and changing instrumentalities. It is also plausible that a reinforcing effect exists here. A mass organization is a prerequisite for, or at least it bolsters the chances of, a successful mobilizing strategy. In turn, a successful mobilization strategy enhances the raison d'etre of the mass membership.

Declining Functionality and Rival Structures

The functional perspective in analyzing party decline has also attached major significance to rival structures or forms of political organization which impair the political party's capacity to perform its functions and even take over party functions. In Epstein's view, however, the dwindling functions of parties do not constitute a problem inasmuch as structuring the vote is the minimum and essential function performed by parties. In fact, he argues that the electoral function may be accomplished more easily if parties are freed from certain functions, more specifically from the representation of interests and what he calls the 'programmatic' function.

By contrast, other writers regard the loss or weakening of functions as a source of party decay. If, for example, parties do not fulfil the function of representing interests, and if other structures are perceived as better equipped to carry out this function, people will transfer their allegiances from parties to these alternative forms of political organization. The growing political role of interest organizations and the rise of ad hoc action groups and single issue movements in many Western countries have increased speculations that these forms of political organization pose a threat to parties. It is hypothesized that the political party is in the process of being superseded or bypassed by these structures in fulfilling the functions of activating people and representing their interests and aspirations.

This hypothesis has been advanced in Scandinavia, and at first glance three

developments appear to enhance its plausibility. First, membership in organizations is very commonplace. According to survey data, membership rates in Scandinavia are among the highest in the Western countries (Pestofl 1977, 65). In the late 1970s around 90 per cent of the adult population claimed membership in at least one organization in Denmark (Johansen 1980, 66-8) and in Sweden (Westerståhl & Johansson 1981, 53f). The Norwegian figure in the mid-1970s was roughly 70 per cent (NOU 1982: 3, p. 105). The Danish and Swedish percentages also suggest that an increase in organizational membership has occurred since the early 19705. Swedish data further indicate an upswing in membership activity in the form of attending and participating in meetings (Westerståhl & Johansson 1981, 50, 55f, 133).

Second, the major economic organizations have been successively incorporated into the policy process. Already in the mid-1960s the now classic statement 'votes count but resources decide' (Rokkan 1966, 105) pointed tc the mounting importance of the organizational-corporate channel of government decision-making in Norway, and the statement would have been equall> applicable to Sweden and Denmark. Since the mid-1960s the major economic organizations have continued to consolidate their influence. These organizations have also been extremely successful in recruiting members, thus strengthening claims to represent the interests of their members.

Third, single issue movements and ad hoc action groups have become ar increasingly familiar feature in the three countries since the mid-1960$ (Gidlund 1978; Olsen & Sætren 1980, 36). The movement against the war ir Vietnam in the late 1960s and the anti-EEC movements in Norway anc Denmark and the People's Campaign against Nuclear Power in Sweder during the 1970s can be mentioned to illustrate the point. In addition, a myriac of ad hoc action groups and organized protests have crystallized, and ir quantitative terms they constitute the mainstay of single issue politics (Olser & Sætren 1980, 46; Andersen 1980, 106).

Estimates of the extent of participation in action groups activities anc single issue politics during the latter half of the 1970s exist for Denmark anc Norway. Approximately one-third of the Danish population (ages 18 to 70; and nearly one-half of the Norwegian population (15 and older) are estimatec to have participated in these activities — which range from taking part ir demonstrations, protest meetings, and political strikes to writing to a news paper or signing a petition (Andersen 1980, 106f; Olsen & Sætren 1980, 46) Scattered pieces of data suggest that even a larger proportion of Swedes have taken part in these sorts of activities than have the Norwegians and the Danes.3

In short, other forms of mass political organization appear to be increas
ingly important in Scandinavian political life. The question is: what are th<
consequences for the political parties and especially the Social Democrats'

First, and more generally, are these other forms of political organization rival or hostile forms in the sense that they threaten the parties' function of mobilization? Second, and more specifically, the electoral difficulties of the Social Democrats coincided with the growing prominence of alternative forms of political action. Is there a relationship here?

Action Groups and Single Issue Movements

A major impression from available Scandinavian data is that action group activities are neither hostile to the parties nor is this form of political action superseding the parties. Instead, data from all three countries show that persons engaged in party politics have a greater propensity to participate in this form of political action than those who are not involved. Voters participate more than persons who abstain from voting, and party members more than voters. Norwegian and Danish data similarly reveal that active party members participated more than nominal members, and that a large proportion of party activists were simultaneously ad hoc activists (Westerståhl & Johansson 1981, 43; Andersen 1980, 128, 131; Olsen & Sætren 1980, 48, 83).

Nor are action group activities isolated from party politics. In fact, in the Norwegian case, ad hoc activities to a substantial degree appear to be an extension of routine politics and partisan conflicts.⁴ Taking a closer look at action group activities in different issue areas, a Norwegian study found that participation in various campaigns was usually related to party preference. For example, participation in environmentalist actions was associated with voting for parties with a pro-ecology stance, and participation in anti-abortion campaigns was strongly correlated with voting for a different set of parties. The main exception when the political constellation in single issue politics differed sharply from the party alignment in parliamentary politics was local issues and regional allocation of resources. Not surprisingly, campaigns focusing on local and regional concerns mobilized support which cut across all party lines. Finally, it was often the 'losers' in the regular channels of politics who utilized this form of political action (Olsen & Sætren 1980, Chapters 3 & 4).

Further, an expansion of ad hoc political activities does not inevitably entail a contraction of party activities. In Sweden participation in action groups and other ad hoc activities so far has not occurred at the expense of involvement in party activities. The percentage of party activists in the late 1970s was roughly the same as in the late 1960s (Westerståhl & Johansson 1981, 41, 52f). Norwegian party membership statistics also fail to indicate an attrition in party involvement during the past decade. In absolute terms, the number of Norwegians who were party members in the late 1970s had increased since 1969. In relative terms, as measured by the ratio of party members to voters in 1969 and 1977, the proportion of party members

remained virtually unchanged (Svåsand 1981, 15-16; Statistisk årbok 1981, 373). In Denmark, however, party enrollment waned in the 19705. But it is debatable whether this decline is directly attributable to ad hoc participation; the decline had set in long before the upsurge in involvement in single issue politics.

Finally, participation in action group activities does not necessarily signify a rejection of parties as a form of political action by ad hoc participants who are not party members. On the contrary, a Danish study found that those involved in ad hoc activities were relatively favourably disposed toward joining a political party. A larger proportion of ad hoc participants could contemplate becoming a party member than non-participants (Andersen 1980,

In conclusion, the evidence cited here, although preliminary and disparate, gives cause to question the notion that action groups and single issue movements are hostile forms of political action which threaten the mobilization function of parties. Instead, judging from the Swedish and Norwegian cases, an expansion of other forms of political activation can occur without a contraction of party activity. On the basis of Norwegian evidence, ad hoc activities in many instances represent an extension of party politics. Partisans are often in the vanguard of single issue politics. The Danish data further suggest that participation in this form of political action might eventually lead to party membership.

Thus far the discussion has centred on single issue politics and parties in general. What sorts of variations are found among the parties? And in particular what is the relationship between Social Democratic partisanship and participation in this form of political action? Unfortunately, data on partisanship and single issue politics are not available for all three countries, but data do exist for Norway and Denmark.

Contrary to the popular image that persons involved in this form of political action are one-sidedly recruited from the far left, we find that ad hoc participants in both countries come from the entire political spectrum. Moreover, in Norway non-socialist partisanship among ad hoc participants outweighed socialist party preferences. (The term socialist refers to the Social Democrats and the parties to their left). Among Danish participants, however, socialist partisanship was predominant.

Despite this contrast, there are notable similarities in the pattern of Social Democratic sympathizers' participation in action group activities in the two countries — both with regard to the proportion of ad hoc participants who are Social Democratic voters and the proportion of Social Democratic voters who participate in this form of political action. Roughly one-fourth of the participants in (and organizers of) ad hoc activities in Norway and Denmark reported Social Democratic preferences. Thus Social Democratic sympathizers comprised the largest contingent among ad hoc participants and activists in

terms of supporters of a single party. An examination of the proportion of various parties' sympathizers who have participated in ad hoc activities reveals that approximately one-third of the Social Democratic backers in both countries were ad hoc participants. In Denmark this proportion is quite similar to that of the sympathizers of several other parties, although a number of parties had higher participation rates. In Norway (where the figure for DNA voters was 38 per cent) Social Democratic supporters were clearly the least inclined to participate in such activities (Olsen & Sætren 1980, 48; Andersen 1980, 123-9).

Looking more closely at the various participation rates in ad hoc activities among party sympathizers we can try to gauge the effect of ad hoc participation on a party's efforts to mobilize support. In accordance with the rival structure hypothesis, specifically that action groups and single issue movements are supplanting the parties, one would expect that high ad hoc participation rates among party sympathizers would be correlated with declining membership and support. Ironically, then, if these assumptions were correct, the moderate participation rate among Social Democratic voters might be interpreted as a blessing in disguise.

An empirical test which compares the rate of ad hoc participation among voters for the Norwegian parties and membership trends neither confirms the expectations of the rival structure hypothesis nor yields a coherent pattern. Of the parties with the highest participation rates, the Christian People's party (65 per cent) and the Centre party (62 per cent) followed divergent membership patterns during the 1970s.⁵ The enrollment of the Centre party declined, but only slightly. The Christian People's party, contrary to the assumptions of the rival structure hypothesis, nearly doubled its membership during the decade. The parties whose sympathizers had moderate participation rates also display dissimilar trends. The Liberals' membership decreased drastically, although part of the decline was no doubt related to the party rift on the EEC issue. In contrast, the DNA, with the lowest participation rate (and also plagued by internal dissension on the EEC issue) attracted new members from the mid-1970s onwards (Olsen & Sætren 1980, 48; Svåsand 1981, 13-18). Thus, even close scrutiny of the level of ad hoc participation among sympathizers of various parties fails to substantiate the notion that this form of political action is intrinsically negative for the parties.

In summation, available Scandinavian evidence suggests that action groups and single issue movements are forms of political organization which are supplementary, rather than hostile, to political parties, including the Social Democrats. This, of course, is not to argue that the content of single issue politics, the issues which action groups and movements seek to put on the political agenda, does not have repercussions — both positive and negative — for individual parties (cf. Andersen 1980, 133). The content of single

issue politics during the 19705, in many instances, was scarcely propitious for the Scandinavian Social Democratic parties. For example, the anti-EEC movements in Denmark and especially in Norway siphoned off support from the Social Democrats and clearly contributed to the electoral difficulties of the DNA in the 1973 election (Valen & Martinussen 1977). Similarly, the issues raised by Swedish environmentalist groups and the People's Campaign against Nuclear Power appear to have had a decisive influence on the SAP's defeat in the 1976 election and its subsequent failure to regain power in 1979 (Petersson 1977, 198, 200-01; Korpi 1981, 147f, 155-61). In short, the content and not the form of single issue politics tells us more about the Social Democrats' fading electoral fortunes in the 19705.

A final observation concerns the linkages between single issue politics and party politics. Where relatively strong parties exist, as in Norway and Sweden, the likelihood increases that action group activities will represent an extension of party politics to some degree. It is also more likely that there will be a greater overlapping of party and ad hoc activists. For example, in Norway over two-fifths of the vanguard of single issue politics were party members, whereas in Denmark around one-fifth were party members (Olsen & Sætren 1980, 48, 82-84; Andersen 1980, 125, 128). Differences in the organizational strength of parties may be crucial in shaping the linkages between party politics and single issue politics.

Interest Organizations

The rival structure hypothesis also claims that interest organizations pose a threat to political parties. A major thrust of the argument is that the direct involvement of interest organizations in government decision-making jeopardizes the parties' function of policy formulation, but an examination of this argument would lead us far beyond the scope of this article.

More relevant to the discussion here is the idea that representation of interests by organizations has undermined the representative function of parties, thus causing disaffection and adversely affecting the mobilization potential of parties. In Scandinavia this idea has perhaps found clearest expression in the discussion of the rise of corporatist politics in Denmark (Dahlerup et al. 1975; Jarlov & Kristensen 1978, 66f). Of the three countries, Denmark also provides the most pronounced case of membership and participation in organizations outstripping that in parties. Certain trends and events in Denmark during the 1970s gave further credibility to this variant of the rival structure hypothesis, such as the contrast between increasing membership rates in organizations and declining party enrollment, vacillations in the electorate, along with voter dissatisfaction with the established parties, especially in the protest election of 1973.

Other trends and data, however, scarcely lend support to the notion that
organizations have weakened the mobilizing capacity of the parties, or that

the Danes have transferred their sympathies from parties to organizations. First, electoral participation during the 1970s was consistently high — between 86 and 89 per cent — and thus one of the highest participation rates among the Western countries. The high turnout rates of the 1970s do not indicate widespread disaffection among voters with parties per se or a serious demobilization of the electorate.⁶ Second, during the decade the proportion of party identifiers among voters remained fairly constant, except in 1973 when party identification fell considerably (Borre 1981, 8f)- The similar levels of party identification in the Danish electorate before and after the protest election of 1973 show that there was no enduring deterioration in the voters' attachments to parties. Third, data on the Danish electorate's affective and cognitive orientations towards parties and organizations do not reveal that organizations have superseded parties (Nielsen 1982; Damgaard & Kristensen 1981, 20). Thus a closer look at the Danish case offers little confirmation thai: strong organizations adversely affect the mobilization function of the parties.

In fact, there is much evidence that the opposite is true, and that strong organizations enhance the mobilizing capacity of parties. Numerous studies document a positive correlation between organizational affiliation and various forms of involvement in party politics: voting, campaigning activities, party membership, participation in party meetings, contacts with party officials and holding party office. For example, Verba, Nie and Kim's cross-national study on political participation (1978) revealed that the impact of organizational affiliation on voting and campaign activity largely replicated the impact of party affiliation. Originally, the authors had expected to find substantial differences between the impact of organizations and parties. They assumed that party affiliation would have an appreciably greater impact on participation in elections and campaigning activity than organizational membership. Instead, they found that both promoted participation, whereas non-affiliation discouraged participation in party politics.

Similar findings exist for Scandinavia. Data show that both Swedish and Norwegian households with no organizational affiliations had the highest percentages who failed to vote and lacked party preferences (Pestoff 1977, 130). Conversely, party activity was associated with active involvement in organizations (Oscarsson 1976, 196). In fact, the correlation was so strong in Norway that a recent study concluded that organizations and parties were not independent channels of political influence (NOU 1982:3, 126). A major Danish investigation of political participation, in examining the correlations between various kinds of participation, also noted a relationship between participation in organizations and parties (Andersen, Buksti & Eliassen 1980, 213). Generally, then, affiliation and active participation in organizations reinforce, or at a minimum correlate with, involvement in party politics.

A number of features characterizing the pattern of organizational life in

the three countries undoubtedly strengthen the effects of organizational membership on involvement in party politics. Of prime significance is the manner in which organizations are politicized. In Scandinavia, several organizations have linkages with parties and can be classified as party supportive organizations (Valen, in Pestoff 1977, 44) which provide a party with resources (financial contributions, members, leaders) or as party integrative organizations which not only provide resources but also transmit and/or reinforce party norms among members of the organization (Pestoff 1977, 45f; cf. Oscarsson 1976, 195ff). The clearest example of the latter is, of course, the tråde unions (LO) which are avowedly Social Democratic. In addition, farmers' organizations back agrarian-based parties, and business associations are supportive of non-socialist parties. In Norway temperance and religious associations with ties to the Christian People's party also qualify as party supportive organizations.

Secondly, large proportions of Scandinavians are members of politicized organizations, and this assuredly contributes to the overall impact of organization affiliation on party involvement. A third feature is that the organizations generally mirror the major social cleavages in the countries, and many politicized organizations are characterized by a distinctive base or social homogeneity. Finally, Swedish and Norwegian data suggest that crosspressures resulting from multiple memberships are not very strong (Pestoff 1977, 128).

These features assume added importance for the Social Democratic parties. As party integrative organizations which transmit or reinforce party norms, the LO unions have been relatively successful in imbuing members with Social Democratic norms, although the degree of effectiveness varies among the three countries. Using party choice as a gauge of assimilation of Social Democratic norms, we find that in the late 1970s among the LO members who voted, the proportion voting Social Democratic ranged between a little over one-half (54 per cent in Denmark in 1979) and two-thirds (68 per cent in Sweden in 1979) in the three countries (Nielsen 1982, 46; Holmberg 1981, 339; Valen 1981, 113).

Equally important, the LO unions organize and penetrate large sections of the electorate. In the three countries LO enrollment equalled at least one-fourth of the electorate at the close of the decade. More precisely, members of the Norwegian LO constituted 25 per cent of the eligible voters in 1977, whereas the Danish LO membership comprised 32 per cent of the electorate and the Swedish membership 35 per cent in 1979.⁷ The proportions of the electorate with union affiliations are boosted when household membership is included. Norwegian and Swedish survey data reveal that respondents in households where a member of the family belongs to a LO union were also very prone to vote Social Democratic (Pestoff 1977, 168, 116). In short, the

organization-party nexus is vital to the Social Democratic parties, and the mobilization function is shared by the LO unions and the party. Without the unions, the mobilizing capacity of the Social Democratic parties would be seriously, if not irreparably, reduced.

Accordingly, the variant of the rival structure hypothesis arguing that interest organizations pose a threat to parties stands out as singularly misguided in the case of the Social Democratic parties. Nor do we find much evidence that it is otherwise correct. In all three countries membership and active participation in organizations are correlated with various forms of involvement in party politics, indicating that organizational affiliation has reinforcing, rather than hostile, effects for political parties. Not even in the Danish case where party membership has declined at the same time as organizational affiliation has grown, do data tapping voters' attitudes towards and perceptions of parties and organizations show that interest organizations have displaced parties.

Summary and Conclusions

This article has examined two major theses concerning the declining functionality of political parties: firstly, Epstein's argument that changing instrumentalities, especially new techniques in political communication and party financing, have fundamentally altered the functional necessity of the mass party and, secondly, two variants of the rival structure hypothesis.

The application of Epstein's thesis to the Scandinavian Social Democratic parties discloses that the experiences of the SAP confute his argument. Obviously, the deviant case of the SAP does not 'disprove' his thesis but an inspection of this case js important in several respects. First, the case of the SAP demonstrates that technological advances in political communication need not ineluctably spell the demise of the mass membership party or its importance to the electoral function. This case underlines the importance and variety of possible strategic responses that a party can make. Secondly, an examination of this case helps to identify relevant variables which contribute to sustaining mass membership. In addition to the more commonly acknowledged factor of collective affiliation, patterns of electoral mobilization and party competition as well as the option of a mobilizing strategy and the party's perception of the option appear to be an essential part of the explanation behind the exception of the Swedish Social Democrats.

Moreover, the whole argument of the declining functionality of the mass party underestimates the significance of organizational strength to the electoral function of parties. This de-emphasis, in part, seems to derive from Epstein's view of the functions of parties and particularly the electoral function. He not only has a minimalist view of the functions of parties in that the electoral

function constitutes the minimum function. He also conceives of the electoral function in minimal terms. It consists of structuring the vote — or in Epstein's words — 'the imposition of an order or pattern enabling voters to choose candidates according to their labels' (77). This may merely involve providing the label itself and acquainting the voters with it. With this definition, the variable of organizational strength becomes quite insignificant. A broader definition, one which includes not just providing labels, but also structuring political opinions and mobilizing voters, would restore the organizational variable to its proper place in the analysis of the electoral function.

In the case of the rival structure hypothesis, the preceding discussion has presented substantial empirical evidence which is hardly in accord with the thesis — either with regard to action groups and single issue movements or interest organizations. Action groups and single issue movements in Scandinavia appear to represent a form of political action which is essentially supplementary, rather than hostile, to political parties, including the Social Democrats. Most evidence indicates that organizational affiliation has a reinforcing effect on the parties' efforts to mobilize support. For the Social Democratic parties, membership in organizations, particularly unions, assumes special significance. The organization-party nexus is a major cornerstone in the strength of the Social Democratic parties.

The discussion here also suggests a fundamental weakness in the assumptions of the functional line of reasoning. Implicit in the functional perspective is a fixation with the monopolization or specialization of functions. It is assumed that the perfect situation for a particular^ structure, including political parties, is a monopoly over a specific function or set of functions. The breakdown of this monopoly and the loss of functions are viewed as a process of decay.

On the matter of the loss of functions Epstein's views differ. He does not regard the loss of functions as a sign of decay as long as the minimum electoral function remains intact. He further argues that benefits ought to accrue from a loss of functions, primarily by eliminating goal conflicts or difficulties in simultaneously fulfilling two or more functions. Rather than a sign of decay, he views dwindling functions as a mark of modernity inasmuch as it involves a specialization of functions. In fact, the view that a greater degree of differentiation and specialization of functions between parties and interest organizations represents modernity is an important part of his claim that US parties are more 'developed' or 'modern' than European mass membership parties.

Emphasis on the specialization and monopoly of functions tends to block consideration of the possibility that shared functions, as distinct from monopolized or specialized functions, could benefit a structure. The Scandinavian experience underlines that, in the case of Social Democratic parties,

the organization-party nexus is crucial and shared functions may entail
sizable advantages to political parties.

---

NOTES

1. The description here differs from Elvander's analysis which covers developments through 1974. He presents a picture of similar decline for the SD and DNA. Membership trends of the DNA in the 1970s were characterized by sharp fluctuations. The 1970s witnessed the most dramatic exodus of members in the party's recent history. From 156,000 members in 1972, enrollment plummeted to roughly 132,000 in 1974. By the late 19705, however, enrollment had returned to nearly the same level as the early 1970s (Svåsand).

2. For a discussion of the Swedish Social Democrats' mobilization strategy during the 1948 election, see Sainsbury 1980.

3. According to election survey data, 60% of the Swedish electorate in the late 1970s had signed a petition and a little over 25% reported having taken part in demonstrations (Holmberg & Nordlöf 1982, 52). These percentages are higher than those reported for Norway and Denmark.

4. Of course, this is not to suggest that the Norwegian findings can be generalized to Denmark and Sweden. At this early stage in the research on ad hoc activities and single issue politics, it is impossible to say to what extent the results of the Norwegian investigation would be replicated in the other two countries.

5. The sympathizers of the Socialist Left party had the highest ad hoc participation rate (67 per cent) but membership data for this party are lacking. The rates here refer to the participation among party voters. Rates for party members yield a somewhat different rank order, with the Liberals ranking next highest after the Socialist Left party.

6. The electoral participation rate in the 1981 election dropped to 83 per cent, which was the lowest turnout since the early 19505.

7. These percentages are calculated on the basis of LO membership statistics and official election statistics (Statistisk årbok 1982, 70, 376; Statistisk årbog 1980, 314, 348; Statistisk årsbok 1980, 249,-422). The figures are probably a bit too high since LO members include foreign citizens without the right to vote in national elections.

REFERENCES

Andersen, J.G. 1980. 'Deltagelse i græsrodsaktioner', in E. Damgaard (ed.), Folkets veje
i dansk politik. Copenhagen: Schultz.

Andersen, J.G., Buksti, J. & Eliassen, K. 1980. 'Deltagelsesmønstre og politisk mistillid', in
E. Damgaard (ed.), Folkets veje i dansk politik. Copenhagen: Schultz.

Borre, O. 1981. 'Fragmentation and Realignment: Electoral Volatility in Denmark 1971-79',
Paper, ECPR Workshop on Electoral Volatility, Lancaster, England.

Dahlerup, D., Jarlov, C, Johansen, L.N. & Kristensen, O. 1975. 'Korporatismebegrebet og
studiet af samspillet mellem politiske institutioner', Økonomi og Politik 49, 317-44.

Damgaard, E. & Kristensen, O. 1981. 'Party Government under Pressure', Paper, the 6th
Nordic Political Science Congress, Turku, Finland.

Einhorn, E. & Logue, J. 1982. Welfare States in Hard Times. Kent, Ohio: Popular Press.

Gidlund, J. 1978. Aktionsgrupper och lokala partier: temporära politiska organisationer i
Sverige 1965-75. Lund: LiberLäromedel/Gleerup.

Holmberg, S. 1981. Svenska väljare. Stockholm: LiberFörlag.

Holmberg, S. & Nordlöf, H. 1982. Valundersökning 1979, Teknisk rapport. Stockholm:
Statistiska centralbyrån.

Jarlov, C. & Kristensen, O. 1978. 'Electoral Mobility and Social Change in Denmark',
Scandinavian Political Studies 1 (New Series), 61-78.

Johansen, L.N. 1980. 'Organisationsdeltagelse', in E. Damgaard (ed.), Folkets veje i dansk
politik. Copenhagen: Schultz.

Johansson, S. 1971. Politiska resurser. Stockholm: Allmänna Förlaget.

Korpi, W. 1981. Den demokratiska klasskampen. Svensk politik i jämförande perspektiv.
Stockholm: Tiden.

Kronvall, K. 1975. Politisk masskommunikation i ett flerpartisystem. Lund: Studentlitteratur.

Kuhnle, S. 1975. Patterns of Social and Political Mobilization: A Historical Analysis of the
Nordic Countries. London: Sage.

Nielsen, H.J. 1982. 'Electoral Politics and the Corporate System: The Question of Support',
Scandinavian Political Studies 5 (New Series) 43-65.

NOU 1982:3, Maktutredningen, Sluttrapport.

Olsen, J.P. & Sætren, H. 1980. Aksjoner og demokrati. Bergen: Universitetsforlaget.

Pedersen, M.N. 1981. 'Denmark: The Breakdown of a "Working Multiparty System'", Paper
the 6th Nordic Political Science Congress, Turku, Finland.

Pestoff, V. 1977. Voluntary Associations and Nordic Party Systems. Stockholm: University
of Stockholm, Department of Political Science.

Petersson, O. 1977. Väljarna och valet 1976. Stockholm: Statistiska centralbyrån.

Rokkan, S. 1966. 'Norway: Numerical Democracy and Corporate Pluralism', in R. Dahl
(ed.), Political Oppositions in Western Democracies. New Haven: Yale University Press.

Sainsbury, D. 1980. Swedish Social Democratic Ideology and Electoral Politics 1944-1948.
Stockholm: Almqvist & Wiksell International.

Sainsbury, D. 1982a. 'Structural and Functional Hypotheses of Party Decline', Paper, ECPR
Workshop on Problems of Party Government, Aarhus, Denmark.

Sainsbury, D. 1982b. 'The "Dilemma of Proletarian Electoralism" and Scandinavian Social
Democratic Parties', Paper, Annual Meeting of the Swedish Political Science Association,
Gothenburg/Skagen.

Scase, R. 1977. 'Social Democracy in Sweden', in W. Paterson and A. Thomas (eds.), Social
Democratic Parties in Western Europe. London: Croom Helm.

Svåsand, L. 1981. 'Membership Trends in Norwegian Political Parties', Paper, ECPR
Workshop on Party Adaption, Lancaster, England.

Thomas, A. 1977. 'Social Democracy in Denmark' in W. Paterson and A. Thomas (eds.),
Social Democratic Parties in Western Europe. London: Croom Helm.

Torgersen, U. 1962. 'The Trend toward Political Consensus: The Case of Norway', Acta
Sociologica 6, 159-72.

Valen, H. 1981. Valg og politikk — et samfunn i endring. Oslo: NKS-Forlaget.

Valen, H. & Martinussen, W. 1977. 'Electoral Trends and Foreign Politics in Norway: The
1973 Storting Election and the EEC Issue', in K. Cerny (ed.), Scandinavia at the Polls.
Washington, D.C.: American Enterprise Institute.

Verba, S., Nie, N. & Kim, J. 1978. Participation and Political Equality. Cambridge: Cambridge
University Press.

Westerståhl, J. & Johansson, F. 1981. Medborgarna och kommunen. Stockholm: Liber