University of Uppsala

1. Introduction

The economic sphere is beginning to attract more attention from political scientists. The functionalistic and system-analytical theories' narrow delimitation of the political system - and their tendency to designate the political and economic systems as separate subsystems within the social system - are beginning to be regarded as irritating and repressive. The problems which despite their fundamental political importance have so far been neglected are those of allocation of economic resources and of distribution of costs and benefits. These are at last attracting increasing attention in political science research. It is symptomatic that William C. Mitchell, who began as a political sociologist and wrote a standard work on Talcott Parsons, now advocates a shift in the focus of political theory, from political sociology to political economy. One of the points to which Mitchell draws attention is the odd fact that the political scientists have ignored tax policy: while there is a wealth of literature, some of it empirical, dealing with financial theory and taxation in terms of economics, hardly anything has been written in the field of political science on either the decision-making and the political process in tax questions or the citizens' attitudes to taxation.¹

The policies of taxation may be regarded as the most significant of all domestic issues. It is of critical importance for the political system's distribution, among the different social groups, of the scarce values produced by the society. It is a matter of constant concern to both political parties and most interest organizations. It continually gives rise to new attitudes toward various subsidiary questions. There is, therefore, every reason for a political scientist to study tax policy - irrespective of his attitude toward the theoretical question of determining the boundary between political and economic systems. Indeed tax policy seems - for the reasons

---

* This article is a summary of parts of my book Svensk skattepolitik 1945-1970. En studie i partiers och organisationers funktioner, Scheduled for publication in late 1972.

* This article is a summary of parts of my book Svensk skattepolitik 1945-1970. En studie i partiers och organisationers funktioner, Scheduled for publication in late 1972.

stated above - to be not the least appropriate subject of research if we are to
analyze functions which parties and interest organizations are assumed to fulfill
in a Western political system.

The main purpose of my investigation of tax policy in Sweden since World War 11, the chief results of which are here presented, is to illustrate the distribution of functions and the interplay between parties and organizations. The theoretical point of departure is thus structural-functional - taking into account the limitations this approach imposes on the possibilities of explaining causes and of analyzing political and economic power relations - because no systematic comparative analysis of the political activities of parties and organizations, intended to clarify the basic functions of a political system of Western type, has as yet been made, either in Swedish or, as far as is known, in the political science research of any other country.

2. The Theoretical Framework

I found a natural starting point for my investigation in Gabriel Almond's wellknown concepts 'interest articulation' and 'interest aggregation' and the normative generalizations he uses when he illustrates his functionalistic scheme of analysis with empirical examples. To be sure, Almond points out that there is no sharp dividing line between the two functions and the structures which fulfill them. Yet he also says that in the modern, highly developed political systems it is primarily the 'associational interest groups', such as trade unions and business organizations, which articulate interests, while the aggregation function is mainly performed by political parties. He expresses his belief in this state of affairs as fitting by approving of the British political system, where he thinks this distribution of roles exists. Conversely, he describes France (before De Gaulle) and Italy in harshly critical terms as examples of political systems where the two functions are unfortunately combined and where there is no effective interest aggregation in the party system.²

The terminology and the views of Almond and his colleagues are fairly generally accepted as regards the functions of articulation and aggregation and their respective distribution among organizations and parties in Western democracies. The criticism voiced against Almond's scheme of functional analysis was not applied to this particular point.³ On the contrary, Almond's concepts and the thought behind them - that in a properly functioning democratic system the organizations are chiefly responsible for the articulation of particular group interests, while the parties combine them to a collective opinion on the basis of a view of politics as a whole - proved useful both for comparative studies of different political systems and for analysis of the system in a specific country. Examples can be found in a series of textbooks and general works in political science.₄

Nevertheless neither Almond nor any other author has tried to make these concepts

precise and useful for empirical investigations. It would seem difficult, if not impossible, to do so, as the concepts are so vague and so abstract. Thus the functionalists of Almond's school make no distinction between the aggregation function and the parties' use of group-oriented standpoints and arguments in order to win votes among different groups. But it is by no means clear that the parties' adjustment to group interests in the parliamentary arena is identical to their combination of promises to different groups in order to win votes in the electoral arena: separate concepts should be used for the two arenas. In my investigation I used the term 'group appeals' for the electoral arena. Another problem involves the empirical use of the very distinction between articulation and aggregation. This is obviously a question of a difference in degree. At what stage, for example, can an organization's articulation of group interests be said to go over to aggregation? How is one to compare the degree of articulation and aggregation in parties and organizations when there is no term covering the intermediate points on the scale from articulation of specific group interests to an all-inclusive combination of interests? Another peculiarity in Almond is that articulation of interests seems to mean only that demands and wants from the surrounding society are passively recorded and later integrated into the political system. Almond does not discuss the possibility that organizations, parties, and other structures may play an active part in interest articulation by anticipating the citizens' wants, arousing needs, and creating opinion.

The same criticism may to some extent be raised against David Easton's conceptual framework. Easton's conversion of 'wants' to 'demands' has its closest equivalent in what Almond calls interest articulation, while the reduction and combination of demands into a smaller number of distinct and coherent alternatives for political action stand for the aggregation function. Both these functions are performed by 'gate-keepers' (parties, organizations, etc.). Thus Easton uses the concept of 'gate-keeping' in two senses, which is advantageous in that the functions in reality are difficult to distinguish. On the other hand, though, it means that the concept is still wider and vaguer than Almond's terms. For the most part Easton too neglects the anticipation of wants. To be sure, he mentions en passant that the interest groups - which in modern industrialized systems are 'specific structures for the conversion of wants' - may anticipate wants and initiate demands.⁵ Yet the point of his theoretical construction lies in the influx of demands and in the necessity of reducing their number by gate-keeping so that the system is not overloaded to the breaking point. This perspective allows little scope for anticipation of wants and the eliciting of demands - a fact which plays an important role in a modern 'service democracy' of, for instance, Sweden's type.

Guided by Almond's and Easton's theoretical works we can make two general
assumptions which may be empirically tested:

1. In a well-functioning democracy of Western type, the function of formulating
the special interests of the different groups is mainly performed by the interest
organizations.

5 Scandinavian Political Studies

2. In a political system of this type the parties carry the main burden of reducing
and combining the demands of the organizations into alternatives of political
action on the basis of a view of politics as a whole.

These two assumptions may be supplemented by a third which can be derived
from Almond:6

3. The distribution of roles may be less marked in a multi-party than in a twoparty
system: the parties tend to represent special interests, and more adjustment
of group interests is made between parties than within them.

These assumptions, which seem to be mere common sense and which often occur in textbooks on political science, guided me in my investigation of tax policy. Their validity has been tested by questions and hypotheses derived from them. The third assumption., of course, can only be completely tested in a comparative investigation of two or more political systems. A partial answer, however, may be deduced from a comparison of the Swedish political parties. If, for example, the three small non-socialist parties should prove to represent as wide a combination of group interests as do the Social Democrats, who are in power - with a percentage of the electorate varying in the period from 1945 to 1970 between 42 percent and 51 percent - this would seem to contradict the thesis of a difference between a multi-party and a two-party system.

Some concepts will be defined before the results of the investigation are presented. First, the word 'function' designates only the 'articulating' and aggregating' activities which organizations and parties are assumed to perform within the political system. No other functions will be examined.

That which is articulated and aggregated is described as group interests. This means standpoints and arguments which may be related to the economic interests of the different groups in the society. The choice of wording denotes that the investigation is concerned not with the conversion of wants to demands but only with the views which organizations and parties publicly present to the decision-makers in the form of remiss answers to commission proposals, manifestos, party motions, etc. (these documents and also the debate in the press have been subjected to systematic content analysis). This also means that the first stage in the aggregation process', the adjustment of opinions within organizations and parties, will not be discussed.

It should be clear from my discussion of Almond and Easton that I find myself unable to use their concepts. To denote the oscillation between Almond's articulation and aggregation functions, and to avoid so comprehensive a designation as Easton's gate-keeping, I shall instead use the following expressions: 'formulation of pure group interests', 'formulation of interest combinations', and 'formulation of general interests'. The term formulation also includes anticipation of group interests as well as the answers of organizations and parties to expressed wants. The

expression pure group interests means that an organization or party adopts a standpoint concerned with the material benefit of a certain group and argues more or less openly on the basis of this group's interests. The term interest combination, of course, designates a coordination of a number of group interests. The concept general interests is reserved for those cases - which do in fact appear from time to tune in tax debates - where the standpoint has no apparent connection with any definite group or groups and where the arguments are based wholly on general ideas of public interest, such as economic balance. This is one way to formulate interests on the basis of a view of society as a whole, which could be described in Almond's terminology as 'aggregation on the highest level' - an aggregation of interests which is so inclusive and diffuse that the group interests are superseded. The expression 'general interests' is used in my analytical language as a technical term, with no evaluating or normative connotation, to denote the opposite of a precise formulation of a particular group interest. Ture group interests' and general interests' should therefore be regarded as extremes on a continuum, and between them there is a series of points at which a number of group interests are combined. I indicate by this usage that I do not accept a concept of an 'objective' or 'real' public interest to which the actors' views on tax policy may be related. I also make a distinction between formulation of general interests and the actors' often repeated attempts to promote their manifest group interests or combinations of interests in the guise of the public interest, or their assertions that these are compatible with the public interest.

It is, of course, worthwhile to consider this line of argument - indeed the idea that there is a public interest plays an important part in politics.⁷ Yet the purpose of this analysis is to penetrate the group interests and to try to relate standpoints and arguments to them. The use of 'general interests' thereby becomes a kind of remainder: the term is reserved for those cases in which standpoints and arguments cannot be clearly related to group interests. Usually, however, it is only a question of a more or less far-reaching approximation to what is described as general interests.

Another vital problem of definition concerns my use of the term conflict dimension. This concept has more or less the same meaning as 'cleavage', namely a deep and lasting antagonism between the parties and within the electorate which is related to a specific issue or consists of a pattern of several issues which may be more or less open or latent.⁸ I distinguish three conflict dimensions in Swedish tax policy: 1) government finance, which refers to the contrast between heavy pressure of national taxes versus general reduction of taxation, and which generally coincides with the attitude toward the size of the public sector; 2) distribution policy, which concerns the division of the burden of taxation among the different classes of society; and 3) equality. The latter conflict dimension refers to the special problem of the design of family taxation, i.e. joint or individual taxation of married couples, justified among other things by arguments concerning equally of the sexes and similar treatment of married and single persons. These three analytical instruments are used for a comparative analysis of the activity of parties and

organizations in taxation policy. Their positions - the word denotes a summary of standpoints and arguments - in the different conflict dimensions may be fixed and related to each other, and shifts in position determined. For practical reasons I shall here use the expressions 'to the right of and 'to the left of, although this does not mean that I shall try lo assign the positions in the different conflict dimensions according to a uniform right-left scale. Finally, I should add that I shall also use the term 'politicization', which means that a question has by a party or parties been brought into the electoral arena and become an issue between the parties.⁹ Different degrees of politicizadon and tendencies to depoliticization of subsidiary tax questions may be observed within the conflict dimensions, but, on the whole, tax policy is always politicized.

3. The Decision-Making Processes

Space does not permit a detailed account of the development of Swedish tax policy since 1945. Yet a number of main features must be mentioned briefly. Altogether twelve different decision-making processes have been investigated, most of them concerned with partial reforms. Three large reforms, involving fundamental changes in the tax system, have been carried through. In 1945 the property tax and death duties were tightened up, as was the taxation of large incomes, while the income tax was reduced for the majority of wage earners. This reform was, therefore, clearly concerned with distribution policy, and was passed by a Social Democratic government - supported on the mam points only by the Communists and the Swedish Trade Union Federation (Landsorganisationen, LO) - after one of the fiercest political battles which has ever occurred in Sweden. This strife gave rise to successful right-wing propaganda in favor of a general lowering of taxes and of a reduction in the public sector, which left its mark on the political debate of the 1950'5. The Social Democrats - who in 1951-1957 formed a coalition government with the non-socialist Farmers' Party - retreated. In 1952 and 1956 there were general reductions in the income tax. In 1952 parties and organizations were more or less in agreement (although at odds on the questions of the property tax and death duties), but in 1956 there was discord in the conflict dimension of distribution policy: thus the LO rejected the suggested tax reduction.

At the end of the fifties, however, the Social Democratic government began a counter-offensive to the right-wing propaganda for tax reductions. One of its results was the second great tax reform. In 1959, despite a united opposition, the government forced through a general indirect tax in the form of a sales tax. The sales tax, which was accepted in the following year by the three non-socialist parties, provided the state with the resources for a strong reform policy during the sixties. There was a powerful expansion in the public sector, financed to some extent by gradual increases in the sales tax. Partial reforms in 1961 and 1965 brought about a redistribution of the tax burden from direct to indirect taxation. These reforms were passed with, to a great extent, the support of the non-socialist parties. Tax

policy was, therefore, to some degree, depoliticized: only the Communists held fast
to their opposition in principle to increases in indirect taxation.

The third great reform, carried through in 1970, comprised continued redistribution from direct to indirect taxation (the sales tax had been superseded in 1968 by a value-added tax, in accordance with the wishes of big business and the opposition). Nevertheless the reform was concerned with distribution policy, which was related to the ideas of equality brought to the fore in political debate in Sweden during the latter half of the sixties. The income tax was reduced only for lower and average incomes, while the tax was increased for persons with high incomes or large fortunes. The biggest innovation was the introduction of individual taxation of married couples. This reform was intended to bring about equality between the sexes and to induce married women to take up gainful employment: this was instigated by the large wage-earner organizations among others. In contrast to those of 1947 and 1959 the reform of 1970 was passed more or less unanimously. The government motion was supported to a great extent by the Liberal Party (Folkpartiet) and the Center Party (Centerpartiet) - after the Farmers' Party left the government in 1957, the party changed its name.

To what extent do parties and organizations participate in the decision-making processes in tax policy? How is the initiative promoting tax reforms divided between parties and organizations? To what extent are the riksdag's decisions in tax questions in harmony with the demands put forward by parties and organizations? The answers to these questions will give a general picture of the distribution of roles between the government, the parties in opposition, and the interest organizations.

It appears that the parties (with the exception of the Communists), through their representatives, played major roles in the twelve Royal commissions on tax policy which I investigated. The organizations were not represented on most of the commissions. The government - which in the Swedish system appoints all commissions of investigation - thus regarded tax policy as the domain of the political parties. On the other hand almost all the organizations investigated were active when it came to remiss answers to the commissions' proposals. This is the chief method by which the organizations can influence tax policy. As a rule the actual initiative for tax reforms comes from the government. The only significant exception was the reform of 1952, which was undoubtedly instigated by the non-socialist opposition. The organizations are of small importance as initiators. The same applies to their influence on the making of decisions in the riksdag. Most of the organizations have their own way only in details - insofar as their wishes do not coincide with the government's intentions. The government is more likely than any other actor to realize its intentions in riksdag decisions. The Liberal and Conservative (Hoger) Parties were relatively successful, however, in the reform of 1952. Like the Center Party they also had some success in questions of detail around 1960. On the whole, however, the government - usually supported by the LO - has dominated Swedish tax policy since 1945.

4. The Formulation of Interests

Are there any general differences between parties and organizations as regards the formulation of interests? Are there any differences in this respect from party to party and between different types of organizations? These are the main questions in my investigation. The latter question will be discussed first. Some hypotheses have been deduced from the answers to the questions, hypotheses which are related to the three theses inspired by Almond.

The investigation shows that as regards tax policy the Swedish parties are chiefly characterized by their formulation of various interest combinations. Yet in some cases formulation of pure group interest occurs. The Communists, who are the smallest party in the riksdag, and who have a higher percentage of industrial workers among their voters than any other party, speak for the interests of the lower income group with more consistency and more bias than the other parties. The Farmers' Party had in the forties and fifties a small and homogeneous group of voters, and the party's concern was at that time narrowly restricted to the interests of the agrarian population. From around 1960 the party widened its register to cover a broad range of group interests, based mainly on the low income groups. At the same time the party had remarkable success in widening and increasing its electorate. The Conservatives experienced a similar success in the fifties: the party's profile of interests became less socially exclusive than it had been during the great struggle over the reform of 1947; it won a series of victories at elections; and its electorate came from a somewhat wider social base. Thus a correlation emerges between the size and degree of homogeneity of a party's core of voters and its tendency to formulate pure group interests. The correlation may be expressed in the form of two hypotheses as follows:

1. The smaller and more homogeneous the group of potential voters, the more
the party will tend to formulate pure group interests.

2. The larger and more heterogeneous the group of potential voters, the more
the party will tend to formulate combinations of interests.

(It should be observed that the correlation only expresses a static relation. It may, however, be reversed. For instance, an already small, homogeneous group of supporters may, should they find themselves continually reduced in members, provoke an effort by a party to increase the group of potential voters by broadening the profile of interests. The development of the Center Party is an example of this.)

An obvious difference emerges between the Social Democrats and the nonsocialist parties, in that the latter often function as channels for specific and very limited business interests (trade, small enterprises). This is particularly true in questions concerning the technical shape of indirect taxation, because organizations of tradesmen and small enterprises have very close contacts with the sitting members of the non-socialist parties. The Social Democrats, on the other hand, formulate broad combinations of the interests of workers and of lower salaried

employees, which at times - especially in 1959 - approach formulation of general
interests. The observation may be generalized in the form of a third hypothesis:

3. The closer the relations between the parties and organizations with limited
spheres of interest, the more they will tend to formulate pure group interests

The same type of correlation may be observed in the organizations as in the parties. In almost all the tax questions investigated, the big mass organizations with heterogeneous memberships revealed a strong tendency to formulate general interests. This was particularly true of the Swedish Cooperative Union's Wholesale Society (Kooperativa Forbundet, KF), the organization with the most heterogeneous composition. The Consumers' interest', as it is represented by the KF, has the character of a diffuse 'general interest'. Indeed, on some occasions a formulation of general interests appeared in the KF's comments on tax proposals. Even the big organizations, the LO and the Central Organization of Salaried Employees in Sweden (Tjanstemannens Centralorganisation, TCO), together with the joint organization on tax questions consisting of big business, the banks, and the insurance companies, the so-called Business and Industrial Tax Committee (Naringslivets skattedelegation), have a tendency to consider the tax problems from the point of view of the general public. They do not usually relate their views to the group interests of their members but have a more comprehensive, diffuse formulation of interests. On the other hand, the small and more homogeneous organizations of the wholesale trade, the retail trade, and small enterprises and agriculture, together with the Central Organization of Swedish Professional Workers (Sveriges Akademikers Centralorganisation, SACO) - which is smaller and more homogeneous than the TCO - are clearly inclined to formulate pure group interest. The following hypotheses may be put forward:

4. The smaller and more homogeneous the organizations, the more they will tend
to formulate pure group interests.

5. The larger and more heterogeneous the organizations, the more they will tend
to formulate general interests.

The analyses of the twelve decision processes investigated yielded a clear result: the differences from party to party and between different types of organizations are so great that there are no general differences between parties and organizations. Shades of difference may be observed, however, by means of a discussion of the goals of parties and organizations, and of their positions in relation to each other in the various conflict dimensions.

The goals of the parties in a multi-party system are - to use the terminology of Gunnar Sjoblom - to realize as far as possible the ideas in their programs (program realization), to win as many votes as possible in the electoral arena (vote maximation), to attain the greatest possible influence in the parliamentary arena (maximation of parliamentary influence), and finally to secure internal harmony (party cohesion). Program realization is the supreme goal, while the other

three may be regarded as subsidiary, or strategic, goals. The points of interest here are primarily the strategic goals of vote maximation and maximation of parliamentary influence. These two goals are of course connected: a party tries to win votes in order to come into power, either alone or in coalition with another party or parties. Yet conflicting aims may also arise. For example, parties which are engaged in fierce competition for the same groups of electors encounter difficulties in forming a coalition government with each other.¹⁰

In the Swedish system the goals of the interest organizations are of a different nature. They do not take part in elections; they seek no share in the ruling power. Their supreme goal is program realization in the sense of maximation of their members' long-term interests, Their most important strategic goal may be stated as follows: to formulate the interests of their members in such a way that these win the maximum consideration of the government and the riksdag. They, like the parties, also have the strategic goal of internal cohesion. For smaller and more homogeneous organizations, 'interest maximation' leads first of all to formulation of pure group interests, since this is adjudged to be the most effective (the most plausible and informative) means of promoting and protecting the members' interests in the face of the government. The larger and more heterogeneous organizations, on the other hand, tend to maximize their members' interests by formulating them more in terms of general interests. The motives for this strategic assessment would seem to differ in different organizations (this will not be discussed here).¹¹ The point at issue is instead that organizations seldom formulate interest combinations. Their formulation of interests usually occurs at the extremes on the scale of pure group interest-general interests.

This is the point of emergence of the great difference between the organizations and the parties. If they are to win as many votes as possible the parties must combine the interests of a number of groups. This applies especially to the nonsocialist parties, which, compared with the Social Democrats, have a heterogeneous body of potential voters; to a great extent the three non-socialist parties have been competing for the same voters.¹² In some situations, however, when the opposition's propaganda in favor of tax reductions was regarded as a dangerous obstacle to vote maximation, even the Social Democratic Party has taken on interest combinations in order to win voters outside their core group, the working class. The government's concession to the opposition in connection with the reform of 1952 is the clearest example. Only the Communist Party seems to have no need for interest combinations, for such deviations from the purely low-income group line would not benefit them in any way, either in the electoral or the parliamentary arena.

It may be appropriate at this point to insert the following problem into our reasoning: What positions do the parties adopt in different conflict dimensions in relation to the organizations which represent the groups whose votes the parties are trying to win, or to keep?

There is here an obvious trend: in all the decision processes except three one
or more of the parties adopted a more moderate position than the 'related' organizations.

In the three exceptions, special circumstances prevailed so that no general conclusions could be drawn. In the majority of cases it is in the conflict dimension of distribution policy that the parties' positions lie nearer to the middle than do the organizations': in 1947, 1952, and 1956 the LO took up positions (to the left' of the government; the Business and Industrial Tax Committee and certain other industrial organizations stood 'to the right' of the Liberal and Conservative Parties in 1947, 1952, 1956, and 1970; and on three occasions the agricultural organizations went farther 'to the right' than the Farmers'Party/Center Party. The same pattern can be seen in 1952 and 1956 in the conflict dimension of government finance. We mention only the great tax reforms here, but similar tendencies emerge in certain minor questions. In the equality dimension, on the question of family taxation in 1964-1970, the TCO and the SACO adopted a more radical position than the Conservatives, while their views approximately coincided with those of the Center, the Liberal, and the Social Democratic Parties. In their comments on an important commission proposal in 1964 the SACO and the TCO took up positions in the conflict dimension of distribution policy that were far 'to the right' of those which the non-socialist parties were to adopt in connection with the partial reform of 1965: they wished to go much farther than the parties in the question of reducing the income tax for the middle income groups.

Why does this pattern emerge in the relations between the democratic parties and the organizations which are closest to them? The answer is partly implicit in the question as far as the parties are concerned. The parties are constantly trying to retain the votes of their core groups and at the same time they are seeking supporters in nearby groups of electors. The Social Democrats are chiefly concerned with preserving and strengthening their position of power by mobilizing the votes of the working class, but from the end of the forties until-the middle of the fifties they felt compelled by the opposition to compete for the votes of the middle groups. Their positions on tax questions were therefore somewhat 'to the right' of the LO's. During the sixties there was no such tension between the party and the LO, partly because both in tax policy and other political questions the government veered to the left from 1965 onward.

The Center Party succeeded - except in the unfavorable election in 1956 - in keeping their share of the votes of the agrarian population: it was between 61 percent and 66 percent throughout the sixties. Yet since the proportion of farmers in the population was falling rapidly, the party could only survive by appealing to other groups. The Center Party's many-sided output of group-related standpoints and arguments in connection with the strife over the sales tax in 1959 is to be seen in this perspective, as is its consistent concentration on the low-income groups during the sixties. While the percentage of farmers among the Center Party's supporters fell from 77 percent in 1956, 57 percent in 1960, 48 percent in 1964 to 29 percent in 1968, the number of industrial workers increased from 1 to 2 percent in 1956 to 13 percent even in 1960 and rose in the following general elections to the second chamber to 22 and 26 percent respectively. The small businessmen to whom the Center Party earnestly appealed from the end of the fifties

onward constituted 10-11 percent of its supporters throughout the sixties, compared with 2 percent in 1956. The change in the composition of the group voting for the Center Party is - so far as we can judge from available electoral statistics - the only great change in the social groups represented by the various parties to occur in the postwar period. In comparison with this development the shifts in the electorates of the Conservative and Liberal Parties are of minor interest; they depend to a great extent on the fact that approximately the same groups of voters went from one party to the other and back again.₁₃ To return to the question of the positions adopted by the Center Party on tax questions, these were bound to be less dependent on those of the agricultural organizations, simply because the party was appealing to other groups, although still able to rely on the continued support of the farmers. The action taken by the Center Party as regards tax policy after the end of the fifties provides unmistakable evidence of this state of affairs.

The Conservative and Liberal Parties were evidently unwilling to risk their goal of vote maximation by adopting as extreme positions as did the Business and Industrial Tax Committee and the smaller industrial organizations at times. They had to appeal to a wider range of potential voters. This applies particularly to the Liberal Party, since its members are more evenly distributed among the occupational groups than are those of the Conservative Party. The strategic goal of maximation of parliamentary Influence would seem to have restrained the Conservatives during the sixties. The Conservatives endeavored to adopt positions which were moderate on the whole and did not diverge too much from the 'middle positions' taken up by the Center and Liberal Parties. Thus the Conservatives tried to smooth the way to a reliable non-socialist government alternative and to the acquisition of ruling power.

The tendency on the part of the parties (with the exception of the Communists) to formulate interest combinations and to take up relatively moderate positions on the main questions in the parliamentary arena corresponds to their efforts to include in their propaganda to potential voters 'broad' group appeals such as lowincome groups', 'middle-income groups', and 'families with children'. This was clearly proved by the systematic content analysis of the press debate, radio and television programs, and certain brochures.

If we reverse the question of the positions of the actors in relation to each other and consider the relationships between organizations and parties from the organizations' point of view, we find a natural explanation as to why they adopt more extreme positions than the parties 'akin' to them. The organizations can formulate the interests of their members without the restrictions imposed on the parties by the strategic goals in the electoral and parliamentary arenas. An interest organization need not calculate with voles won or lost or with parliamentary constellations. Its problems instead comprise deciding how the message can be expressed to have the maximum effect on the hearer, taking the divergent views within the organization into consideration, and in certain cases coordinating its views with those of other organizations, etc. This type of deliberation does not normally lead to formulation

of interest combinations - for the simple reason that the organization does not need to consider primarily the interests of other groups. Either it formulates a pure group interest without considering what positions the parties have adopted or may be expected to adopt, or else it represents so large a section of the population or - as in the case of the Business and Industrial Tax Committee

- a sector so central in the country's economic life that it naturally rises above group interests and formulates its position more in terms of general interests. Even the large and heterogeneous organizations often disregard the attitudes of the parties, and they too may adopt positions which are more extreme than those of parties 'close' to them. The latter does not apply to the KF - no party can be said to have a close affinity to the cooperative movement - and it only applied once to the TCO, in 1964. The TCO's answer to the commission's proposal of 1964 was indeed far 'to the right' of the non-socialist parties' positions in 1965, but the circumstances surrounding the tax reform of 1965 were so unusual that no conclusions can be drawn from the case of the TCO in 1964.

We are left with the LO and the Business and Industrial Tax Committee. These alone among the big organizations tend both to formulate general interests on some occasions and to adopt more extreme positions than parties 'close' to them. This is not by chance. The LO and the Business and Industrial Tax Committee stand out as the two great opposite poles in Sweden's economic life. The trade union movement stands against the collective of big business. The interests of the two blocs in the field of tax policy are at times expressed in more ideological terms and in more extreme standpoints than are to be found in the Social Democratic and the Conservative/Liberal Parties. This happened especially in 1947, but to some extent also in 1952 and 1956; it also happened to a limited degree in 1964-1965 so far as the Business and Industrial Tax Committee was concerned, and in 1970. It is, therefore, in situations in which tax questions are fairly or very strongly politicized that the LO and the Business and Industrial Tax Committee take their positions out on the wings. They put forward their opposite interpretations of 'the public interest', unimpeded by the strategic and parliamentary considerations which restrain the parties. The LO acts as a goad on the government and as a guardian of the ideology during a period when the Social Democrats are carrying on a defensive campaign against an aggressive right-wing opposition; this was particularly clear in 1947.

The agrarian organizations took up somewhat more extreme positions than 'their' party, the Farmers' Party, in two tax questions during the fifties, the decade when tax policy was fairly strongly politicized. Moreover all the organizations (except the LO) took up positions on the wings in relation to the non-socialist parties during a brief period of political strife concerning the commission proposal of 1964. Taken together these observations apparently provide a good foundation for the following hypothesis:

6. The more politicized the tax policy, the more the organizations tend to adopt
extreme positions in relation to their 'related' parties.

At this point it is appropriate to ask whether or not there is any connection between the degree of politicization and possible differences between the parties' and organizations' lines of argument. It seems reasonable to assume that the organizations' arguments are more technical and more neutral than those of the parties, especially in slightly politicized questions, while their arguments become more and more political and more like those of the parties in decision processes in which party political divisions are strong. Yet the investigation confirms only the second part of this assumption. It is quite clear that the organizations use technical and more or less neutral arguments in questions which are only slightly politicized. But so too do the parties, on the whole. On the other hand, in questions regarded as politicized the arguments of both parties and organizations incorporate a heavier political emphasis. The following hypothesis may be formulated:

7. The more politicized the tax policy, the more political are the arguments of
the organizations.

If this hypothesis - and the closely related hypothesis 6 - can be confirmed by other empirical investigations, as seems probable, then it is proven that there is an intricate interplay between organizations and parties in a democratic political system of the Swedish type. The organizations closely follow the shifts in the struggle of party politics. Their positions are affected by the degree of politicization. Their standpoints and their arguments are sometimes used by the parties as weapons in the political struggle. In strongly politicized issues, the organizations can speak their minds in a more emancipated and radical fashion than their related' parties.

What is said here implies a not unimportant modification of the distribution of roles between parties and organizations which is usually taken for granted, namely that the parties struggle to attain vote maximation in the electoral arena and maximation of parliamentary influence, while the organizations - at least those indifferent to party politics - seek only to promote their members' interests without side-glances at the parties. The neutral organizations often emphasize their independence of party politics. This is, of course, correct insofar as they generally take sides in politicized interest questions without allowing themselves to be affected by the parties' known or assumed standpoints. Nevertheless, they are greatly influenced by the general political situation - the degree of tension between the parties, the tone of party political arguments, conditions of parliamentary power* tendencies to depoliticization of an issue, etc. The organizations may not be bound by the party standpoints. Yet they are by no means free of the influence exercised by the very game of politics. The parties, moreover, are often decisively influenced by the organizations' views when they make their calculations for vote maximation.

5. Summary and Conclusions

At the beginning of this article two general assumptions were put forward, based
on the theoretical works of Almond and of Easton. These were intended to guide

our investigation: their validity was to be tested by questions relevant to them. This process is now complete in that the questions have been answered and certain hypotheses formulated. The results of this testing will be summarized in a closing discussion of the two central assumptions; the third assumption, concerning differences between two- and multi-party systems, will be mentioned more en passant.

The investigation provided no support for the two generalized assumptions (pages 65f above). Reality is far more complicated. In the generalizations no distinction is made between different kinds of parties and between different types of organizations. My results indicate that this theory is untenable, at least as regards the political system in Sweden. Certain parties have a purely 'articulating' function; certain organizations go so far in their reduction and coordination of special interests that they tend, to a greater degree than any party, to formulate general interests. The most striking relationships between attributes of parties and of organizations and their ways of performing their functions are expressed in hypotheses 1-5.

The third assumption, however, which may also be derived from Almond, involves an important modification of the two main hypotheses. It states that the assumed distribution of roles between parties and organizations is less marked in a multi-party than in a two-party system, because in the former case the parties tend to represent special interests; group interests, therefore, are coordinated by means of compromises between the parties in parliament rather than by internal agreements within the parties as in the two-party system. On the whole this is in accordance with the picture given by the investigation of the Swedish multi-party system. The non-socialist parties have a tendency to formulate pure group interests, so that in some situations they are forced to coordinate their standpoints at the parliamentary level. Thus the Liberal and Center Parties collaborated closely in the riksdag during the latter half of the sixties. On some occasions for example they submitted joint motions on tax questions. To be sure the difference between government and opposition diminished somewhat during the sixties. The Social Democrats were more faithful to the interests of their core group, while the Center Party changed from a narrow 'class party' to a broad 'popular party'. Nevertheless, the non-socialist parties still serve as channels for the special business interests, side by side with the organizations concerned. In a two-party system this special 'interest articulation' would pass through a filter of internal coordination in the non-socialist party. The non-socialist interest combination would probably be more extensive and inclusive than is that of each individual opposition party at present.

Naturally Almond is right in saying that interest combinations are peculiar to the parties of a democratic political system of the Western type: the investigation confirms this view and shows that the explanation thereof is to be found in the parties' strategic goals. Nevertheless, Almond ignores almost completely the fact that certain organizations - the large and heterogeneous top organizations - may go much farther than do the parties as regards the coordination of group interests and formulation of general interests. To be sure, he does point out that European

organizations of this type carry out a far-reaching aggregation of special interests and present 'policy alternatives' to the parties and the government.¹⁴ However, he does not seem to assign ;any great significance to this accurate observation: he does not allow it to modify his theoretical generalizations. Almond's neglect may be explained by his regarding the parties as a kind of filter between the interest groups and the decision-makers.¹⁵ The parties' main tasks are to screen off certain special interests, to modify and coordinate other interests, and, finally, to propose political alternatives based on alternative interest combinations. The aggregation by the parties takes place after the articulation by the organizations. In Almond - as in Easton - this view is so prevalent that the idea that the parties make up their minds on political questions before the organizations do and that the latter turn directly to the government seems wholly alien.

Yet it is just this which is characteristic of the Swedish system, with its welldeveloped system of commissions and remiss answers. The parties often adopt views at the commission stage which coincide with their final standpoints on proposals, party motions, and parliamentary committee reports. In such cases the parties are but little influenced by the remiss answers of the organizations. The organizations' views serve to reinforce the opinions already adopted by the parties, rather than as a basis for their decisions. There is no aggregation of organization interests at the political party level: organizations and parties formulate their interests in parallel, and largely independently of each other. On important questions the organizations address themselves directly to the government by their remiss answers and informal contacts: they do not go via the parties in the riksdag. These circumstances mean that Almond's normative 'theory' on interest articulation and interest aggregation cannot be applied to the Swedish system. It is not only the terminology which is inappropriate, the underlying idea of a certain distribution of roles between organizations and parties also seems irrelevant. A comparative investigation of the functions of parties and organizations in connection with the formulation of interests would probably show that this is also true of other European democracies besides Sweden.

Almond's and Easton's schemes of analysis are also misleading at a more fundamental level. They seem to be based on an ideal picture of representative democracy, in which all initiatives emanate from the people; the citizens as individuals or in groups put forward their interests or demands, and the task of the rulers is confined to the making of decisions on the basis of the reduced and coordinated demands of the citizens. As suggested in the introductory section this perspective allows little scope for the anticipation or guidance of wants, or for opinion-making on the part of the parties, organizations, and government bodies. Above all, Almond and Easton ignore the government's dominant role as instigator and policymaker in the important domestic problems of the distribution of values and allocation of resources in society. They do not distinguish between the great basic questions of distribution and the many minor current problems; they do not realize that the distribution of initiative and the decision process may be entirely different in the two types of questions. Probably their mental picture of the ideal political

system was unwittingly colored by political life in the USA, where the frequency of spontaneous initiative on the part of the citizens and the activity of interest groups (in the form of lobbying) would seem to be higher, relatively speaking, than in the European democracies - at least in the field of home affairs - and where the political system is less centralized than it is in Europe. Almond's theory of interest articulation may well cover the situation in the USA, and perhaps even that in Britain. It is probable that Easton's analytical scheme of the political system is highly suitable for subsystems (parties, organizations, local authorities, administrative bodies) in Western democracies. Nevertheless, it is definite that their theories can only be applied to the national political systems in Western Europe if they are essentially modified in such a way that the government's leading role is maintained as regards the central problems of distribution policy.

The investigation of Sweden's tax policy provides extraordinarily clear evidence of the dominant role played by the government in the big and constantly topical questions of the extent of the state's resources, and the principles to be applied in the distribution of the burden of taxation over the different citizen groups. The roles of the parties and - even more so - of the organizations are largely confined to answering the proposals initiated or inspired by the government, and to trying to avert too unfavorable a distribution of the burden for their own groups. There may well be reason to ask at this point whether or not the tax policy really is representative of distribution issues in domestic affairs in general. This problem was not discussed at the beginning of this essay; the only assumption made was that the politics of taxation would be the most suitable subject for investigation in a study of the functions of parties and organizations as advocates of group interests.

Tax policy evidently differs in one respect from other complexes of domestic issues. In the conflict dimension of government finance the tax policy is, so to speak, the answer to the demands for reform put forward by parties, organizations, administrative bodies, and individual opinion-makers: it provides the society (the state and the local authorities) with the resources necessary to bring about the reforms. So it is only natural that the government, personified by the chancellor of the exchequer, should dominate national tax policy. Only the government has the power and the general view necessary if all the demands for expenditure are to be weighed against each other and against the total (national and local) pressure of taxation. Parties and organizations, etc. tend to accept the government's judgment in the conflict dimension of national finance. The development in Sweden during the sixties - when the disagreements on tax policy were reduced at the same time as the total tax pressure sharply increased - may be regarded as a confirmation of the views herein expressed. The conflict dimensions of distribution policy and of equality allowed greater scope for initiative and influence from other actors than the government - and even from public opinion. My investigation provides examples of this. Nevertheless even here the government generally takes the lead, by reason of the inescapable relationship between the conflict dimension of government finance and the other two. The parties' and the organizations'

role is to avert rather than instigate and promote. By using one of Easton's concepts in a way he did not intend, the parties and the organizations could be said to act in tax policy as sentries at the gates of the citizens' homes, and not, as Easton had in mind, in front of the chancellor of the exchequer's treasury.

We should add that the picture of the government's dominance and the relatively passive parts played by the opposition and the organizations would be somewhat different if the main emphasis of investigation of tax policy had been on current minor matters such as communications from the industrial organizations, annual non-socialist party motions in the riksdag concerning details of the form of corporation taxation, the abolition of point taxes on certain goods, reductions of the vehicle and petrol taxes, etc. The frequency of initiative 'from below' is probably greater in this type of tax problems than in the big tax questions which are subject to investigation by commissions and the remiss system. Possibly this would be a more appropriate setting for Almond's and Easton's ideas on interest articulation and 'gate-keeping. Nevertheless the study of the treatment of such questions, which was made in connection with the analysis of the decisions on indirect taxation taken in 1959, 1965, and 1968, indicates that even here the government prevails, in that the initiative taken by non-socialist parties and concerned interest organizations does not usually lead to any results. Moreover, it is obviously of greater interest to investigate the decision process in the large central tax problems of distribution policy than in the many minor questions which primarily concern special interest groups.

There is, therefore, evidence that the government's major role is more pronounced in tax policy than in other spheres of home affairs. A government which seeks to realize the reforms called for by the party program, and also to satisfy certain requests for expenditure made by opposition parties and organizations, and which is concerned with maintaining the balance of the country's economy, must keep a firm hold on tax policy and see that it retains the initiative in this central area. The opposition and the organizations are thus obliged to be relatively passive, prepared to keep watch. Where expenditure is concerned, they probably have more opportunities of taking the initiative and forcing measures through. The strategic goals of the parties in opposition - vote maximation and maximation of parliamentary influence - constantly incite them to try to outbid the government's reform policy. The organizations make demands involving government expenditure in their special spheres of interest. Parties and organizations formulate and anticipate the interests of the different categories of citizens more actively in special interest areas than in the field of tax policy. It is possible that as regards expenditure the political system functions in closer harmony with Almond's and Easton's theoretical models than proved to be the case in tax policy. This, however, is only a guess at this juncture. The matter should be investigated by systematic analyses of the distribution of initiative, the participation of the various actors in the decision processes, the correspondence between the decisions and the demands made, the character of the formulation of interests, relations between parties and organizations, etc. These analyses should be undertaken in those sectors of expenditure

where the policy has a certain consistency, so that many observations may be made
over longer periods. Certain of the hypotheses here proposed could provide guidance
and be subjected to tests.

---

NOTES

1. W. C. Mitchell, 'The Shape of Political Theory to Come: From Political Sociology to Political Economy', in S. M. Lipset (ed.), Politics and the Social Sciences, New York: Oxford University Press, 1969, pp. 101-136; J. M. Mitchell and W. C. Mitchell, Political Analysis & Public Policy: An Introduction to Political^ Science, Chicago: Rand McNally, 1969, pp. 202-205.

2. G. A. Almond, 'A Comparative Study of Interest Groups and the Political Process', American Political Science Review 52 (1958), pp. 270-282; G. A. Almond, 'Introduction: A Functional Approach to Comparative Polities', in G. A. Almond and J. S. Coleman (eds.), The Politics of the Developing Areas, Princeton, N.J.: Princeton University Press, 1960, pp. 33—45; G. A. Almond and G. B. Powell, Comparative Politics: A Developmental Approach, Boston: Little, Brown, 1966, pp. 73—127.

3. Cf. R. Dowse, 'A Functionalist's Logic', World Politics 18 (1966), pp. 607-622; E. J. Meehan, Contemporary Political Thought: A Critical Study, Homewood, 111.: The Dorsey Press, 1967, pp. 26, 175; L. Rudebeck, 'Political Development: Towards a Coherent and Relevant Theoretical Formulation of the Concept', Scandinavian Political Studies, Vol. 5, Oslo: Universitetsforlaget, 1970, pp. 32-35.

4. See for example R. E. Jones, The Functional Analysis of Politics, Library of Political Studies, London: Ruthledge & Kegan Paul, 1967, pp. 59-75; R. C. Macridis and R. E. Ward (eds.), Modern Political Systems: Europe, Englewood Cliffs, N.J.: Prentice-Hall, 1968, pp. 18-21; A. Banks and R. Textor, A Cross-Polity Survey, Cambridge, Mass.: MIT Press, 1963, passim; E. Rasmussen, Komparativ Politik 1, Copenhagen: Gyldendal, 1968, pp. 105, 110, 173, 182; L. Rudebeck, 'Det politiska systemet i Sverige', in E. Dahlstrom (ed.), Svensk samhdllsstruktur i sociologisk belysning, Stockholm: Svenska Bokforlaget, 1968, pp. 475-480. In The Little, Brown Series in Comparative Politics, edited by G. A. Almond, J. S. Coleman and L. Pye, several authors use Almond's scheme of analysis. See for instance H. W. Ehrmann, Politics in France, Boston: Little, Brown, 1968, Chaps. VII and VIII.

5. D. Easton, A Systems Analysis of Political Life, New York: John Wiley, 1965, Chaps. 4-6, 9, cit. p. 96.

6. Almond and Coleman (eds.), op.cit., pp. 42-44; Almond and Powell, op.cit., pp. 102—103.

7. Cf. G. Schubert, The Public Interest. A Critique of the Theory of a Political Concept, Glencoe, 111.: The Free Press, 1960, passim. After having shown the subjective and arbitrary use of the term 'the public interest' in American political theory, Schubert recommends the elimination of the concept from political science terminology. An opposite, nearly value objectivist conception is maintained by R. E. Flathman, The Public Interest. An Essay Concerning the Normative Discourse of Politics, New York: John Wiley, 1966. An example of a 'realistic' view of the necessity to use 'the public interest' as a concept for political analysis is E. E. Schattschneider, The Semi-Sovereign People. A Realist's View of Democracy in America, New York: Holt, Rinehart and Winston, 1960. For a summary of the discussion on public interest, see C. J. Friedrich (ed.), The Public Interest, Nomos No. 5, New York: Atherton, 1962.

8. Cf. G. Sjoblom, Party Strategies in a Multiparty System, Lund Political Studies 7, Lund: Studentlitteratur, 1968, pp. 122-123, 169.

9. Ibid.

10. Ibid., pp. 73-95.

11. Cf. N. Elvander, Intresseorganisationerna i dagens Sverige, 2nd rev. ed., Lund: Gleerup. 1969, p. 297.

12. B. Sarlvik, 'Political Stability and Change in the Swedish Electorate', Scandinavian Political Studies, Vol. 1, Oslo: Universitetsforlaget, 1966, pp. 188-222.

NOTES

1. W. C. Mitchell, 'The Shape of Political Theory to Come: From Political Sociology to Political Economy', in S. M. Lipset (ed.), Politics and the Social Sciences, New York: Oxford University Press, 1969, pp. 101-136; J. M. Mitchell and W. C. Mitchell, Political Analysis & Public Policy: An Introduction to Political^ Science, Chicago: Rand McNally, 1969, pp. 202-205.

2. G. A. Almond, 'A Comparative Study of Interest Groups and the Political Process', American Political Science Review 52 (1958), pp. 270-282; G. A. Almond, 'Introduction: A Functional Approach to Comparative Polities', in G. A. Almond and J. S. Coleman (eds.), The Politics of the Developing Areas, Princeton, N.J.: Princeton University Press, 1960, pp. 33—45; G. A. Almond and G. B. Powell, Comparative Politics: A Developmental Approach, Boston: Little, Brown, 1966, pp. 73—127.

3. Cf. R. Dowse, 'A Functionalist's Logic', World Politics 18 (1966), pp. 607-622; E. J. Meehan, Contemporary Political Thought: A Critical Study, Homewood, 111.: The Dorsey Press, 1967, pp. 26, 175; L. Rudebeck, 'Political Development: Towards a Coherent and Relevant Theoretical Formulation of the Concept', Scandinavian Political Studies, Vol. 5, Oslo: Universitetsforlaget, 1970, pp. 32-35.

4. See for example R. E. Jones, The Functional Analysis of Politics, Library of Political Studies, London: Ruthledge & Kegan Paul, 1967, pp. 59-75; R. C. Macridis and R. E. Ward (eds.), Modern Political Systems: Europe, Englewood Cliffs, N.J.: Prentice-Hall, 1968, pp. 18-21; A. Banks and R. Textor, A Cross-Polity Survey, Cambridge, Mass.: MIT Press, 1963, passim; E. Rasmussen, Komparativ Politik 1, Copenhagen: Gyldendal, 1968, pp. 105, 110, 173, 182; L. Rudebeck, 'Det politiska systemet i Sverige', in E. Dahlstrom (ed.), Svensk samhdllsstruktur i sociologisk belysning, Stockholm: Svenska Bokforlaget, 1968, pp. 475-480. In The Little, Brown Series in Comparative Politics, edited by G. A. Almond, J. S. Coleman and L. Pye, several authors use Almond's scheme of analysis. See for instance H. W. Ehrmann, Politics in France, Boston: Little, Brown, 1968, Chaps. VII and VIII.

5. D. Easton, A Systems Analysis of Political Life, New York: John Wiley, 1965, Chaps. 4-6, 9, cit. p. 96.

6. Almond and Coleman (eds.), op.cit., pp. 42-44; Almond and Powell, op.cit., pp. 102—103.

7. Cf. G. Schubert, The Public Interest. A Critique of the Theory of a Political Concept, Glencoe, 111.: The Free Press, 1960, passim. After having shown the subjective and arbitrary use of the term 'the public interest' in American political theory, Schubert recommends the elimination of the concept from political science terminology. An opposite, nearly value objectivist conception is maintained by R. E. Flathman, The Public Interest. An Essay Concerning the Normative Discourse of Politics, New York: John Wiley, 1966. An example of a 'realistic' view of the necessity to use 'the public interest' as a concept for political analysis is E. E. Schattschneider, The Semi-Sovereign People. A Realist's View of Democracy in America, New York: Holt, Rinehart and Winston, 1960. For a summary of the discussion on public interest, see C. J. Friedrich (ed.), The Public Interest, Nomos No. 5, New York: Atherton, 1962.

8. Cf. G. Sjoblom, Party Strategies in a Multiparty System, Lund Political Studies 7, Lund: Studentlitteratur, 1968, pp. 122-123, 169.

9. Ibid.

10. Ibid., pp. 73-95.

11. Cf. N. Elvander, Intresseorganisationerna i dagens Sverige, 2nd rev. ed., Lund: Gleerup. 1969, p. 297.

12. B. Sarlvik, 'Political Stability and Change in the Swedish Electorate', Scandinavian Political Studies, Vol. 1, Oslo: Universitetsforlaget, 1966, pp. 188-222.

---

13. Data on the party electorates are taken from J. Westerståhl and B. Sårlvik, Svensk valrorelse 1956: Arbetsrapport I, Intervjuundersokningen, Goteborg, 1957 (mimeo), and from Riksdagsmannavalen åren 1959—1960, 11, Sveriges Officiella Statistik: Allmånna Val, Stockholm: Statistiska Centralbyrån, 1961, Riksdagsmannavalen åren 1961-1964, 11, Sveriges Officiella Statistik: Allmånna Val, Stockholm: Statistiska Centralbyrån, 1965, and Riksdagsmannavalen åren 1965-68, Del 2, Sveriges Officiella Statistik, Stockholm: Statiska Centralbyrån, 1970; cf. Sårlvik, op.cit.

14. Almond and Coleman (eds.), op.cit., p. 39; Almond and Powell, op.cit., p. 100.

15. Almond and Coleman (eds.), op.cit., p. 37.

13. Data on the party electorates are taken from J. Westerståhl and B. Sårlvik, Svensk valrorelse 1956: Arbetsrapport I, Intervjuundersokningen, Goteborg, 1957 (mimeo), and from Riksdagsmannavalen åren 1959—1960, 11, Sveriges Officiella Statistik: Allmånna Val, Stockholm: Statistiska Centralbyrån, 1961, Riksdagsmannavalen åren 1961-1964, 11, Sveriges Officiella Statistik: Allmånna Val, Stockholm: Statistiska Centralbyrån, 1965, and Riksdagsmannavalen åren 1965-68, Del 2, Sveriges Officiella Statistik, Stockholm: Statiska Centralbyrån, 1970; cf. Sårlvik, op.cit.

14. Almond and Coleman (eds.), op.cit., p. 39; Almond and Powell, op.cit., p. 100.

15. Almond and Coleman (eds.), op.cit., p. 37.