SOU 1999:151

Citizen Participation in European Politics

Citizen Participation in European Politics

Hans Agné Cees van der Eijk Brigid Laffan Britta Lejon Pippa Norris Hermann Schmitt Richard Sinnott

Demokratiutredningens skrift nr 32

SOU och Ds som ingår i 1999 års nummerserie kan köpas från Fakta Info Direkt. För remissutsändningar av SOU och Ds som ingår i 1999 års nummerserie svarar Fakta Info Direkt på uppdrag av Regeringskansliets förvaltningsavdelning.

Beställningsadress: Fakta Info Direkt, Kundservice Box 6430, 113 82 Stockholm Tel: 08-587 671 00, Fax: 08-587 671 71 E-post:


In October 1999 a conference was held in Stockholm on Citizen

Participation in European Politics. The background of the dis-

cussion was a low Swedish turn-out in the Elections to the European Parliament in June. We wanted to give an account of and discuss research on citizen participation in European politics illuminating possible explanations. Special attention was given to the connection between on the one hand the level of participation and on the other hand the role of the institutions and of the mass media.1

During the two days five international and four Swedish scholars contributed by giving lectures consisting of a variety of challenging perspectives and empirical data. In order to stimulate the public debate we are now pleased to publish most of the lectures.

The conference was arranged by The Commission on Democracy in Sweden in cooperation with The Swedish Ministry of Justice and The Committe of Evaluation of the General Election of 1998.

The organisers want to express their deepest gratitude to professor Sören Holmberg for his generous support during the planning of the conference. However, neither professor Holmberg nor the organisers but the authors themselves are finally responsible for the opinions formulated in this book.

Erik Amnå

Principal Secreterary The Commission on Democracy


1 In Swedish the Commission on Democracy also has published other analyses dealing with global and European challenges to Swedish democracy, for instance Demokrati på europeisk nivå? (SOU 1998:124), EU – ett demokratiprojekt? (SOU 1998:145), Bör demo-

kratin avnationaliseras? (SOU 1999:11) and Globaliseringen och demokratin (SOU 1999:56).

Table of Contents

D e m o c r a y w i t h i n t h e E U i n t h e L i g h t o f t h e 1 9 9 9 E u r o p e a n P a r l i a m e n t a r y E l e c t i o n s

Britta Lejon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

W h y S o m e P e o p l e V o t e a n d O t h e r s d o n o t

Cees van der Eijk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 3

E u r o p e a n P a r l i a m e n t E l e c t i o n s . I n s t i t u t i o n s , A t t i t u d e s a n d P a r t i c i p a t i o n

Richard Sinnott. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 7

I n t e g r a t i o n a n d S t a t e h o o d i n t h e E u r o p e a n U n i o n . T h e d i s t r i b u t i o n o f P o w e r s B e t w e e n t h e E U a n d t h e M e m b e r S t a t e s

Brigid Laffan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 5

T h e A r c h i t e c t u r e o f E U I n s t i t u t i o n s a n d C i t i z e n P a r t i c i p a t i o n

Hermann Schmitt. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 7

B l a m i n g t h e M e s s a n g e r ? P o l i t i c a l C o m m u n i c a t i o n s a n d T u r n o u t i n E U E l e c t i o n s

Pippa Norris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 9

T o S h a r e D e m o c r a t i c L e g i t i m a c y B e t w e e n D i f f e r e n t P o l i t i c a l L e v e l s

Hans Agné. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 7


Democracy within the EU in the Light of the 1999 European Parliamentary Elections

Member of Cabinet Britta Lejon, Ministry of Justice

Ladies and gentlemen,

I am happy to welcome you all to the seminar on ”Citizen Participation in European Politics”. This seminar is focused on one of the most important common goals in Europe: To grant that the European Union will be and percieved to be a union of the people of Europe.

We who share that vision are terribly set back by the poor turnout in the elections to European parliament this summer.

A low participation by the citizens is the greatest obstacle in the process of securing peace and prosperity in Europe.


The EU has been given important tasks. They involve guaranteeing peace, ensuring a good environment, creating favourable living conditions, and much besides. We do not consider national decisions to be sufficient in addressing such problems as precarious security, environmental threats, organised international crime, flows of capital that shift jobs from one country to another with lightning speed, to name but a few. We need political decision-

making that can deal powerfully with those problems for which individual nations lack sufficient decision-making authority.

Decisions reached within the European Union have major consequences in nearly all policy areas. EU policies have effects on the work of national parliaments as well as on that of municipalities and county councils. The EU influences the everyday life of people living within and in fact also far outside the borders of the community. Moreover, the influence of the EU in such areas is steadily increasing. The Treaty of Amsterdam has put a number of political issues on the Union agenda which are essentially of an everyday nature, for example consumer policy, social issues, equality between men and women and employment policy.

At the same time we know that decisions and decision-makers in the political sphere must have popular legitimacy and support.

The European elections are the single occasion on which the citizens of Europe can join forces to influence the shape of the EU’s future. Nevertheless only half, 49.9 per cent, of the EU citizens entitled to vote did so in the European elections held this summer. The election turnout in Sweden was 38.8 per cent.

We cannot accept elections in which well below half of the eligible voters participate. There are a number of reasons for concern.

First, we consider that a low election turnout reduces the legitimacy of the institution concerned.

Second, we are afraid that a low turnout in one election will rub off on other elections. That is to say, a low turnout in the European elections might also get citizens into the habit of not voting in national elections.

Third, a high election turnout has its own intrinsic democratic value. It is indicative of a broad political debate within society, in which most people familiarise themselves with current political issues and make their own judgements regarding the political development they would like to see. If few take part in elections the risk is the opposite – that democracy will be diluted and that a representative government will in a sense be reduced to government by a minority.


To reverse the trend towards low electoral turnout we need knowledge. This is why the Swedish Government gave the Committee of Evaluation of the General Election of 1998 the additional task of analysing the results of the 1999 European elections.

It is also why the Commission on Democracy in Sweden, the Committee of Evaluation of the General Election of 1998 and the Ministry of Justice, have invited you to take part in this seminar.


There might be different views on the ideals of democracy. But the absolute majority of people are in full agreement on the key role of general elections in a democracy. The EU is not a nation, it is a very special organisation but it does have a directly elected parliament. However, the authority of this parliament cannot be equated with that of a national parliament.

People in our countries have realised that power in the European Union is not concentrated exclusively in the Parliament. An important reason why the European Parliament does not attract the same interest as a national parliament is undoubtedly that significant power is vested in the national governments.

But the low interest in the European Parliament reflects a low interest in democratically elected institutions generally. From the point of view of democracy a number of negative trends can be seen in society today, both in Sweden and in most of the mature democracies. There is declining confidence in political institutions. Public participation and the confidence of citizens in the ability of the established parties to solve problems and take action is being successively weakened. Public participation in society requires greater efforts and permanent commitment – both of which are decreasing.

As far as I can see, the situation is not dramatic in the present EU member states. Democracy as a system is not threatened – on the contrary. There is considerable popular support for democracy as the best form of governance.

But since public confidence and participation are long-term social phenomena, it is important to heed the warning signs early on.

Beside the low turn-out we also have the results of the recent elections in Austria and Switzerland.

We know that the European Union is facing massive and in part difficult challenges. Enlargement eastwards is one of them.


We must identify which problems are specifically related to the EU. Efforts to make the work of the EU more transparent are an important part of this. Sweden has a long tradition of public access to information. The question of transparency applies to all EU institutions. The EU must be an organisation of its age. Modernising the institutions of the EU is crucial for public participation.

We need transparent and understandable processes. We need budgetary disciplin, accountability, equal opurtunities for men and women , access to documents among other things. But in order to boost turnout at European elections we must also reflect upon what role we would have the EU play in the future.

However, I don’t think it will be enough to implement institutional reforms at the EU-level. The political parties, other NGOs, the mass media etc, will also have to revise the way in which they work.

Many EU-related issues have had a divisive effect on the traditional parties. Nevertheless, I believe that these parties must have the courage to put EU issues on the agenda, partly for the sake of their own future and partly because we cannot allow people’s confidence in politicians to deteriorate still further. A study made in 1998 shows that over 60 per cent of voters consider themselves to be uninformed when it comes to EU-related issues. A mere 2 per cent consider themselves to be very well informed.

Among other things there must be an effort to increase and deepen information to the public about union-decisions and the progress in the fulfilling of them. This is a responsability for the EU-institutions – specially the council – and the member states.

Creating our national democracies took decades. Building European cooperation on a democratic foundation must also be

allowed to take time, perhaps even longer bearing in mind how new the East European democracies are.

After this introduction I look forward to hearing your views and discussing these matters with you over the next two days. In the coming discussions about the future development of the Union the Swedish governments intend to keep focus on popular support and participation. There must be a joint analysis of the reasons for the low election turnout and proposals of measures that might reverse the trend.

Thank you.


Why Some People Vote and Others do not

Cees van der Eijk1

The question of the determinants of participation in elections has occupied political scientists and political practitioners alike, particularly since the introduction of universal suffrage. This question arises from the basic observation that not all people who have the right to vote make use of that right: while some people vote, others do not. Consequently, one is tempted to ask whether this difference originates in other distinctions between these two groups, and if so, what other characteristics are involved.

Before World War II, this question had to be addressed mainly by analysis of percentages of turnout by region, municipality, or even smaller geographical units. Renowned in this respect are the studies of the French sociographer André Siegfried (1913) and the Swedish sociologist Herbert Tingsten (1937). Notwithstanding the insights these analysts gained, this approach has severe limitations, and may even generate fallacious conclusions. The development of survey methods after World War II, and the explosion of practical possibilities for analysing large amounts of survey data with the development of cheap computer technology have allowed the pursuit of the question of why people do or do not vote at the level of the individual voter. This survey approach has greatly increased our understanding of the individual-level determinants of

1 The author is professor at Department of political science and

Amsterdam school of communication research at university of Amsterdam.

electoral participation. Yet, in spite of the tremendous number of survey-based publications, important questions remain unanswered – particularly concerning the determinants of changes in overall levels of turnout. Most recently, the insight has been established that combinations of individual-level and aggregate or system-level information are required in order to make further progress in answering the question of who votes and who does not, but this approach has not yet been applied very much.

In this report, I will focus on the behavior of individual voters, and the reasons why they do or do not employ their right to vote. I will therefore not systematically address the effects of systemic characteristics on the over-all likelihood of voting – although in passing I will occasionally refer to some of the work in that tradition.

This report is divided in three parts. In the first, I review a number of theoretical approaches to the question of individual electoral participation. The relevant concern in this discussion is

not how successful each of these approaches is in explaining differ-

ences between voters, but rather their conceptual and theoretical logic. In the second part of this report, first findings from a recent study of the Swedish electorate will be reported. These derive from the Swedish segment of the European Election Study 1999. The third part consists of reflective comments that link parts 1 and 2, and that contain some informed speculation about causes of the

low turnout in the

most recent Swedish parliamentary elections

and European elections that, at the moment, cannot be subjected to empirical testing because of lack of relevant data.

Part 1: Theoretical and conceptual approaches

The literature on electoral participation contains a number of approaches that differ in terms of the locus of the factors considered to be of causal importance for making people either participate or abstain in elections. One rarely finds these approaches in ideal form, as most theorists and empirical analysts see the benefits of using elements of them in a complementary fashion, but for the purpose of presentation I will discuss them separately in more or less ideal-type form.

The first of these approaches looks for the explanation of differences in electoral participation to the voters involved: different people have different characteristics that help to understand why they behave differently. The second approach looks at the other actors that are of necessity involved in the conduct of elections: the political parties, the candidates and the politicians. They have to reach and mobilize voters and to the extent that they fail to do so this will explain non-voting. The third approach focuses on something else again: the specific context of the election at hand. Not all elections are alike, and from the differences between them one may try to explain why some people vote and others do not. Each of these three approaches will be discussed in some more detail below.

Who votes depends on voters’ characteristics

The logic of this approach rests on the notion that voters should possess certain qualities in order to turn out and vote. In general one can think of the following kinds of qualities that may relate to turnout:

Physical qualities: voting requires a trip to a ballot station, and

therefore that a person is sufficiently healthy and mobile to make this trip, that s/he has sufficient visual and motor skills to read the ballot and handle a pencil or voting-machine, etc. In most western political systems elections have been organized in such a way as to minimize the physical and motor demands that

are required for casting one’s vote, so that the number of people unable to vote for these trivial reasons is minimized. Yet, for some people the activities required to vote will remain too strenuous – the (very) old, chronically ill, and disabled – although it can be expected that their number is quite small in developed democratic systems such as Sweden.

Economic and social resources such as (flexible) time and

money. To the extent that voting is organized in such a way that it costs a lot of time (because of distance to travel to a polling station, because of queues waiting for their turn, etc.) this may deter people from voting, and people with more free time or a more flexible time schedule will be advantaged. This will be particularly so when people’s income is so low that they cannot afford to lose time that is necessary for acquiring their income.2 In large measure, it is believed that these factors play at most a minor role in preventing people from going to the polls in developed western democracies, were it only because of measures taken to minimize these effects (such as long polling hours, possibilities for absentee voting or postal voting, etc.)

Social connectedness. Elections are a collective enterprise, the

results of which are relevant for a society or political community as a whole. Belonging to that community or society is then an important factor that promotes turnout, whereas being estranged or alienated from it may obstruct electoral participation. For this reason, social integration is a factor that impinges on turnout, and at the individual level this relates to the extent that an individual is integrated in networks of relationships that make a person into a member of the community. Although it is difficult to state which concrete characteristics should be taken as indicative for integration, the relevant literature looks at factors such as membership of and being active in non-political


Although not of relevance in most western societies, the consequences

of this logic run sometimes in the opposite direction: voting can be a source of income in conditions of widespread poverty in combination with a practice of vote-buying. In such circumstances, elections operate as a means of redistributing income.

contexts and organizations – including churches, youth groups, sports and recreational organzations, etc. – and at membership of a multi-person household that bestows upon a person responsibility for others.

Relevant cognitive and information-handling skills. Although

often not perceived that way, voting is a highly abstract act, particularly in the context of a mass society. Usually voters do not know the candidates personally, and parties themselves are often abstract entities. Many of the communications directed at voters assume the presence of some kind of prior knowledge or understanding of politics. The larger a voter’s capacity to handle abstract messages and arguments, the larger the store of experience and knowledge s/he possesses, the better s/he will be able to understand the nature of the voting act in general, as well as in the specific election at hand. Although it cannot be assumed that this understanding will necessarily drive people to cast their votes, it is reasonable to expect that lack of understanding will keep people away. A number of more concrete characteristics relate to cognitive and information-handling skills. Education is a primary one: more education (both in duration as in level) increases one’s capacity for handling information, increases cognitive skills, and provides specific information that helps make sense of what an election is about. In addition to formal education, the amount of previously accumulated experiential knowledge and understanding also helps. Because of this, we may expect –

ceteris paribus – older people to have acquired more understan-

ding and hence to vote more frequently than younger ones. Also those who, for whatever reasons, have had direct experiences with parties, politicians, and political processes in general, will possess more of this facilitative (though not sufficient) condition for voting.

Relevant attitudes. In the course of their lives, from childhood

socialization as well as from experiences in later life, voters have acquired attitudes that may either facilitate or impede electoral participation. Such attitudes are referred to in the literature (e.g. Verba and Nie, 1972) as civic attitudes, and they

pertain to generalized (rather than election-specific) feelings such as the following

¾Self-esteem, sense of political competence and sense of political efficacy, all of which refer to whether a voter looks at him/herself as a person who may have an impact on his/her own life and on what is going on in the world – which makes it reasonable to cast a vote – or as someone who is unable to affect to do so – rendering electoral participation a useless activity. ¾Generalized feelings about the importance of politics in general. A view that the political process is largely irrelevant for whatever happens in the world and in one’s own life is an impediment to electoral participation, whereas the opposite view will promote it. ¾Generalized feelings about the openness, malleability and benevolence of the political system all increase the relevance and potentially beneficial effects of voting, whereas feelings of an opposite nature promote a perception of voting as futile or even hazardous.

Civic attitudes, such as those referred to above, cannot be observed directly in terms of overt behavior. Rather, they have to be inferred from responses to items in survey questionnaires that are expected to yield valid indicators for these attitudes.

The approach in which people’s characteristics are looked upon as factors that are important for their participation (or lack thereof) has been very influential in political science since the 1960’s. Its archetype is perhaps the seminal work by Verba, Nie and others (Verba and Nie, 1972; Verba, Nie and Kim, 1978), later followed by many others (see, e.g. Parry et al., 1992). This tradition has yielded something like a standard model in which socio-economic characteristics of people – representing their resources and social connectedness – impinge both directly as well as indirectly – via civic attitudes, that themselves are strongly determined by these same background characteristics – on electoral participation. Most of the comparative work in this tradition has focussed on the

question how societal and institutional circumstances affect the relative strength of the direct, and indirect impact of people’s background on their participation. An institutional framework of, e.g., compulsory voting, takes away most of the direct and the indirect impact of background, as it renders such distinctions largely irrelevant. Other contexts, such as a largely secularized and ”postmodern” society would yield very strong indirect effects of background, and so forth.

In spite of the wealth of insights yielded by this approach, it is notoriously unable to account for variation in turnout across different elections in a single system. After all, most of the factors identified in this approach as important do not change very rapidly, if at all, over time. Yet, in most countries successive elections yield different rates of turnout, sometimes dramatically so.

Who votes depends on mobilization efforts by parties and others

The line of reasoning of this approach focuses not on voters, but on other actors that are involved in elections. These are first and foremost the parties and candidates that vie for the support of the voters. To the extent that parties and politicians put more effort in reaching people and to ”getting out the vote”, more people will be reached, more people will be affected by the persuasive communications from parties aimed at mobilization, and consequently the number of people casting their vote will be higher. By contrast, to the extent that parties succeed in discouraging the supporters of their opponents from voting (e.g., by so-called ”negative campaigning”) they will have the effect of lowering the number of voters.

Aside from parties and politicians, other actors may also be important in the mobilization process. One can think in this respect of organizations such as pressure groups or cause groups that occasionally take part in the electoral struggle by mobilizing their followers, usually in an effort to make them support a candidate or party sympathetic towards their interest or cause. An important example of such behavior in many countries is the effort by labor

unions to get their members or followers to the polls. Less frequently nowadays in western democracies, but important in former times, were the efforts of churches to mobilize their believers in support of a favored party. Even in the absence of overt efforts by formal organizations, a mobilizing influence can emanate from social pressure, particularly when cohesive social groups exist from which individuals derive (part of) their identity. Obviously, such implicit social pressure and the more open influence of organizations will usually reinforce each other. Such group-based mobilization was the cornerstone of the kind of electoral cleavage-politics that existed in most of European democracies until the 1960’s and 1970’s, but that has lost most of its strength since then.3 Finally, the mass media play an important role in the mobilization process. After all, very few voters have direct communications with politicians or parties, and very few will be informed of appeals by parties and organized groups without media.

When looking at whether or not individual people turn out to cast their vote, the logic of this approach implies that those who have been reached by mobilizing agents – or who have been reached more frequently or more intensively – will vote more often than those who were not. The kind of people that are more likely to be reached by mobilizing agents are those that are comparatively easily targetted as recipients of messages and influence attempts, such as members of parties, organized interest groups or cohesive social groups.4

The logic of this approach is quite compelling when we observe that the efforts of parties and others to mobilize voters vary widely across different elections, and that turnout rates across

3 For the genesis and historical roots of cleavage politics, see Lipset and

Rokkan (1967), Rose and Urwin (1970) and Lijphart (1968). For the most wide-ranging and comparative study that documents the decline of cleavage politics – and of the causes of this decline – see Franklin, Mackie and Valen (1992).


Obviously, people wo are ”eager” for information can also easily be

reached, but this refers not only to the efforts of mobilizing agents, but to the interaction of their efforts with qualities of voters (the first of the approaches, discussed above).

those elections often vary concomitantly. The contrast between national parliamentary elections on the one hand and elections to the European Parliament on the other is a case in point. It has been well documented that the mobilizing efforts of parties and media are much smaller in the latter than in the former, as indicated by their spending in European elections only a fraction of the resources they are accustomed to spend in contests for the national parliament.5 Still, this approach too, has its persistently weak side. When controlling for having been reached by parties, and for having been exposed to mobilizing information and communication, one still observes large differences in turnout. Some people evidently go to the polls to cast their vote even without having been prodded to do so, while for others the most intensive barrage of mobilizing efforts remains without effect. To a large extent these differences in behavior coincide with the differences in voters’ characteristics that were identified as important by the previous approach.

Who votes depends on the election-specific context

In this perspective the starting point is the recognition that every election again is a different one, with its own idiosyncracies. The special aspects of one election, that may help to bring some people to the polling station, may be absent in another election, with the effect that some of them will now stay home, while others will be motivated instead to cast their vote. From this perspective, it depends on the aspects that characterize each election in turn who will turn out and who will not.

In order to avoid confusion, it must be emphasized that this tradition comes in two different forms, one analytical, and one narrative in character. The latter comprises a large number of case studies of single elections – often in pleasant narrative and at first sight quite convincing. Sometimes written by political journalists, sometimes by academics, these accounts try to make a specific election intelligible by pointing to the particular character of the


See, e.g., Blumler (1983), Reif (1985), Van der Eijk and Franklin (1996).

contest, the unique set of candidates or parties that was pitted against each other, the particular political events or possibly even scandals, the charismatic character of some of the main actors, the novelty of how the media operated, the unusual economic or social circumstances that surrounded the election, and so forth. From all these aspects, a chronicle is woven that leaves no room for any other election outcome than the one that indeed materialized. Publications such as these are of great value for getting a ”taste” of what was going on at the election in question, to refresh one’s memory in terms of a chronology or of – indeed – the idiosyncrasies of the contest at that time and place. But they are not of great analytical value. Their defect is that the accounts they provide are in no sensible way testable – hence cannot be rejected as inadequate explanations. A gifted story teller can in this way make a convincing account for any election, using only well-established facts. The trouble is, however, that a different selection from the abundance of well-established facts can equally well be combined into a seemingly convincing story, but entirely different in contents and possibly even incompatible with or contradictory to other ones. This problem can only be remedied by an explicit comparative perspective.

The analytical tradition of the ”specific election context” approach is based on comparison. A specific election context can only be sensibly described by explicit comparison. An account that emphasizes, for example, that an incumbent lost an election because of being embroiled in a scandal, requires for testing that other elections are also investigated in which other incumbents were involved in scandals, as well as elections where the incumbents were not, but where their opponents were, or where nobody was. On the basis of Mill’s time-honored logic of concomitant variation (Mill, 1974/75), only such a comparison can reveal whether or not the attribution of certain outcomes to the factor ”incumbent involved in scandal” holds up or has to be rejected as

ad hoc. It will be obvious that such explicit comparison requires

more abstraction in its explanatory terms, or, in the terms of Przeworski and Teune’s (1970) seminal contribution on the topic,

it requires the replacement of proper names by theoretically interesting concepts.

An important set of such theoretically interesting concepts are provided by the general theoretical framework of rational choice, in which behavior is explained by people’s evaluation of the costs and benefits (monetary or otherwise) associated with each of the alternative courses of action that are open to them. Applied to the question who votes and who does not, the answer will be that people will vote if the benefits of that action outweigh the costs, and that they will abstain in the opposite case. In order to avoid the tautology that all action is rational, as it otherwise would not have been undertaken, costs and benefits are usually restricted to people’s economic and political goals or interests. In a somewhat different, but compatible, terminology, voters evaluate ”what is at

stake” in a particular election and, depending on their individual

goals and interests, their perceptions and expectations, may arrive at different conclusions in different elections. In some situations the stakes are sufficiently high to warrant the trouble of going to the polls, in other situations they are not and abstention will be the result. The task of the analyst is to identify the contextual factors that render a specific election more (or conversely, less) consequential than another for specific voters, and thus to explain their turning out (or conversely, not) in these events.

The perspective of what is at stake is particularly suited for shedding light on differences between elections in terms of individual electoral participation (why does the same voter vote in one election, while abstaining in another one, whereas this pattern may be the reverse for another voter) and of aggregate turnout (why is the percentage of turnout higher in one election than in another?). Barring rather infrequent changes in the electoral system and in electoral procedures, we may assume the ”cost” of voting to be rather invariant across elections.6 Consequently, differences in


In some instances this assumption is unwarranted even in the absence of

changes in electoral procedures or the electoral system. Elections may differ in terms of relevant opportunity costs, as will be the case when one is located in a vacation period and another one is not.

individual or aggregate behavior have to be understood from differences in the ”benefit” side, i.e. in differences of what voters perceive to be at stake. These perceptions can be thought of to be affected by a number of factors.

The most important difference in what is at stake in an election relate to the political, and in particular the policy, consequences an election may conceivably generate. If, for example, no programmatic (or other relevant) differences exist between all (or even most) parties, the stakes are less, because the same policy consequences can be expected to occur irrespective of which party wins.7But even if programmatic differences do exist between contending parties or candidates, the stakes will – ceteris paribus – be low when it can not be expected that these will be reflected in actual government policies, dependent on whom wins the election. In the case of European elections, for example, no clearly discernable relation exists between the election result on the one hand and the policies of the EU on the other hand, as the composition of the European Parliament has no consequences for the composition of relevant policy-making institutions such as the European Commission, the Council of Ministers or the European Council. This is in sharp contrast to elections for national parliaments in parliamentary systems, where the election result affects the political ”color” of the administration. This difference may also be expressed as a difference in (the extent and scope of) executive power that is affected by elections.8 Equally detrimental to the feeling that something important is at stake is the voter’s belief that the outcome of the election is a foregone conclusion. Here perceptions of what other


In this, and the remainder of the argument, it is the perception of the

voter that is of relevance, and not the assessment of an analyst, politician or journalist. For the sake of brevity, this will not always be spelled out in detail in the text. Similarly, whatever differences voters perceive to exist between parties, must matter to them, i.e., must relate to the goals, values or policy preferences they hold.


It is particularly this kind of difference that distinguishes so-called first-

order elections (in which executive power is at stake) from second-order ones, where this is not the case. See, e.g., Reif and Schmitt (1980), and Marsh and Franklin (1996).

voters will do are of immediate relevance, but the existence of such perceptions is not unlikely in view of the information voters acquire about others’ preferences, either via informal communications or by way of pre-election opinion polls. When neither sincere nor tactical voting may conceivably affect the election result in a way that matters for a voter, the stakes are low: the election will then always be expected to yield the same result (desired or not).9

Although this approach is very powerful in explaining the differences in electoral participation between (national) parliamentary or presidential elections on the one hand, and local, regional and European elections on the other, it has so far been less successful in identifying the contextual factors that can explain variations in turnout between elections of the same kind, particularly between different first-order elections (see also note 7). What is the reason for – sometimes quite sizeable – differences in the percentage of people that cast their ballot from one national parliamentary election to the next? The factors mentioned above do not seem to vary very much, so other contextual factors must be called upon to help. As a contextual factor, the ”closeness” of the race has often been hypothesized to be important, but the empirical support for this is so far limited at best, while the applicability of this hypothesis is only straightforward in two-party systems.10Hypotheses about yet other contextual factors are, however, rather scarce in the literature. With respect to second-order elections it has been established that an important contextual factor that affects their importance to voters is their location in the cycle

9 One could object to this line of reasoning, as some theorists do, on the

grounds of the infinitesimal small likelihood that a single voter’s ballot will ”swing” the election result in a way that matters to that voter. The reply to this objection is either that voters’ estimates of this probability is grossly exagerated, or that voters act on the belief that others exist with identical preferences, and that collectively their votes do matter. See also Franklin, (1996).


For a first review of the effect of closeness of the race, see Van Egmond, 1999.

of first-order elections,11 but this too, is insufficient to fully account for, e.g., the decline in turnout in the 1999 elections to the European Parliament in most countries of the European Union.

Returning to the question addressed in this report – why some people vote and others do not – the reasoning above leads to the identification of a series of characteristics of individuals that can be expected to affect their likelihood of voting. First of all, it is necessary that voters have political goals or desires; if they do not, no benefits are to be achieved from voting. Second, they must perceive differences between parties in the respects that matter to them.12 In line with the argument above, voters must have the feeling that (executive) power is at stake and that the election result can go different ways. Some of these latter requirements seem pretty tall orders if voters were to develop such knowledge or information entirely on their own. They can, however, make use of various sources from which they can derive information and cues about the nature of the various parties, about what is at stake, and so on. Thus, in this approach, availability of and exposure to such sources is also an important distinguishing characteristic of voters.13

What is the nature of the act of voting?

The three approaches that were discussed above differ in what they see as the nature of the voting act. In addition, yet other interpre-

11 See Franklin, van der Eijk and Oppenhuis (1996), Oppenhuis (1995)

and Franklin (1996).


Obviously, this also implies that they have more or less developed perceptions about parties or candidates whatsoever.


The extent to which the social and political context provides voters with relevant information and cues in this respect is therefore of particular importance. In a comparative study, Granberg and Holmberg (1988) demonstrated that the Swedish context is much more conducive to voters’ acquiring relevant information than that of the United States. In a different study, Van der Brug (1997) showed that a context that provides sufficient relevant cues can overcome informational deficiencies of voters.

tations are often referred to that are not straightforwardly associated with any of these approaches, but that have to be discussed nonetheless.

The last-mentioned approach, that emphasizes the electionspecific context, clearly looks at voting as instrumental behavior, by which political interests, goals, desires and causes are expressed on the ballot. In this view, differences in turnout indicate that not all elections are seen by voters as equally important as means to further these interests and goals, but such differences in turnout are not necessarily worrisome in themselves. Very low turnout, however, indicates that many voters view the election at hand as barely or not at all relevant to their substantive political interests and concerns. If persistent, this constitutes a threat to the notion of representative democratic governance.

A quite different interpretation of what the voting act means for people emphasizes that it allows the expression of identity and identification. In this view, voting is not intrinsically different from other kinds of behavior that can be understood from the perspective of identity politics, and the usage of which cannot be reduced to instrumental motivations and the cost-benefit reasoning implicit therein. This interpretation of the meaning of the voting act is particularly related to the first two approaches to the explanation of (non-)voting discussed above. When party-related identities exist but are not very deeply entrenched, then mobilizing efforts of parties and others are necessary to ”remind” people thereof. When party-related identities are deeply seated in people’s consciousness, then these will propel them to the ballot box, even without mobilizing efforts by others. When parties (or candidates) are not linked to voters’ political identities or identifications, mobilizing efforts appealing to them will be of no avail, and other (instrumental) appeals have to be made. This view on the meaning of voting does not in itself regard low turnout as worrisome. It merely indicates that the number of people who identify (in a psychologically meaningful sense of the word) with specific political groups or organizations (such as parties) is small. This would only be problematic to the extent that one would view such iden-

tifications as prerequisite for social integration, but that would probably constitute a fallacy of the pars pro toto kind.14

A third interpretation of the meaning of the voting act is that it expresses support for the political system as such. In its negative form, it holds that non-voting expresses lack of support, or even rejection of the existing political system. This interpretation is not directly linked to any of the three approaches to understanding (non-)voting, but may – under appropriate circumstances – follow from each. Not having been socialized into support for the existing political order would lead to abstention, as much as being mobilized into an electoral boycott. Even from an instrumental perspective, non-voting may be the choice for those who reject the current system in favor of another that they expect to be advanced by electoral abstentionism. Yet, neither of these hypothetical situations seems to be empirically relevant in any of the established democracies of the European Union.15 In its strong form (non-voting implying active rejection of the existing political regime), this view of the meaning of the voting act seems hardly relevant. In a milder interpretation, non-voting would be regarded as the result of indifference to or disenchantment with the democratic system, a condition that under specific circumstances would indeed bode ill for democratic stability. This interpretation of nonvoting has in recent years become en vogue in political commentary in various countries. In Germany it became known as the phenomenon of Politikverdrossenheit (a term relating to a syndrome of cynicism, distrust, and withdrawal), in the Netherlands


The tremendous decline in such identifications as reported in the literature on party identification on the one hand (see the review in Dalton and Wattenberg, 1993), and in the decline of cleavage politics on the other (e.g., Franklin, Mackie and Valen, 1992), cannot in any convincing manner be linked to social disintegration.


See, in this respect, in particular the various contributions of the Beliefs

in Government project (Kaase and Newton, 1995), where the relationship

between citizens and their states is probed in rather great detail and as much as possible in historically comparative perspective. Most relevant in the context of this report are (references) Biorcio and Mannheimer (1995) and Topf (1995).

as the kloof (the idea of an unbridgeable chasm between voters and the political professionals).16 In spite of the apparent plausibility of such interpretations, empirical support is lacking. Extensive analysis of the data from a recent study of Germany and the Netherlands led to the conclusion that

… neither the declining membership of political parties, nor the possibly declining rate of participation in elections is sufficient ground for assuming a crisis in the relationship between voters and politics, as is implied in the very terminology of Verdrossenheit and

kloof. Rather, these phenomena reflect changes in the position of

parties vis-a-vis other actors in the political realm. (Van der Eijk and Van der Brug, 1998, p. 27).

In its positive form, however, the ”support for the political system” interpretation holds that the act of voting allows the expression of feelings of civic-mindedness, citizen duty, and so forth, feelings that would be commensurate with the notion of diffuse support for the existing political order. This interpretation coincides with the approach that looks at voters’ characteristics (particularly the presence or absence of civic attitudes) in explaining electoral participation.

Forces impinging on the decision to vote or not

The considerations from the preceding sections can be summarized into three different kinds of forces that – in isolation or in combination – impact on the decsion to vote or to abstain:


Coercion, pressure and persuasion. This may exist in the form

of some legal obligation (such as compulsory voting) or the absence thereof.17 In a different form, the expectation of


The debates on Verdrossenheit and kloof have been summarized in Van der Eijk and Van der Brug (1998). A particularly relevant contribution to these discussions is Van Gunsteren and Andeweg (1994).


Apart from legal conditions, little research has been conducted into the existence of other coercive pressures. ”Common wisdom” holds such

positive or negative sanctions (including social approval or being ostracized) will constitute pressure to either cast a ballot or alternatively to stay away from the polls. Most mobilizing appeals in modern campaigns are focussed on persuasion.


Intrinsic satisfaction. This derives from the possibility of

expressing authentic feelings, such as a sense of citizen duty, or identification with groups or parties, etc. Obviously, depending on the kind of feeling, it may be voting as well as nonvoting that yields satisfaction.


Instrumental motivation. In this case the differential of effort

versus benefit, or the assessment of what is at stake determines whether the voter casts a ballot or not.

In explaining why some people vote and others do not, one can assess the power of each of the different approaches mentioned so far, and subsequently one can combine them into a single explanatory model. In the next part of this report, I will analyse survey data pertaining to electoral participation in Sweden at the occasion of the elections to the European Parliament of June 1999. As is often the case in empirical research, the data available from this study do not cover all factors that one can think of theoretically and that have been mentioned in the previous pages. Lack of space as well as empirical insights from other studies result in leaving out of the questionnaire those factors that are expected to be of little explanatory power. The analyses to be presented will therefore neither for each of the various approaches separately, nor for their combination, yield a full account of the question why some people vote and others do not in this specific historical situation. Moreover, when analyzing only a single election, no adequate account can be given of the impact of all those contextual factors

effects to be minor, but it is not impossible that they play a certain role in personal relationships such as those between, e.g., husband and wife.or pressure from socially cohesive groups that can be expected to exist in systems dominated by cleavage politics – a cindition that has mostly vanished in western democracies (as documented by, e.g., Franklin, Mackie and Valen, 1992).

that exist at these elections, that would require explicit comparison with other Swedish elections, with elections in other countries, or both. Such comparative analyses have been conducted for other elections and other countries, however. In the concluding part of this paper I will summarize the findings of those studies and discuss their possible relevance for understanding electoral participation in the Swedish elections for the European Parliament in 1999.

Part 2: Electoral participation in Sweden in June 1999

The analyses reported below were conducted on data from the Swedish part of the European Election Study 1999 (EES99).

The European Election Study 1999

The EES99 is a survey of voters that was executed in the period immediately following the elections to the European Parliament of June 1999. In each of the member states of the European Union (EU) a random sample of voters was interviewed by telephone, using a questionnaire that on average took approximately 30 minutes to administer.18 This study and the questionnaire were designed and organized by an international research group, just as were its counterparts following earlier European elections in 1989 and 1994. In addition to conducting a voter study, the EES enterprise conducts other kinds of research related to the European electoral and representation process, such as studies of elites (MP’s, MEP’s), studies of the manifestoes and election pledges of the various political parties, studies of legislative behavior by MEP’s and their party groups in the European Parliament, and studies of the way in which the mass media reported and commented about these elections and their campaign.

The 1999 EES voter-study group consisted of Sören Holmberg (University of Gothenburg), Hermann Schmitt (University of


The Italian sample was not interviewed by telephone, but by teleinterviewing instead. This sample consists of the members of an existing panel study, who downloaded the questionnaire and completed it on their individual PC’s, to then upload the completed version again. The samples were in all other cases drawn randomly. Unsuccessful attempts – due to not reaching the intended respondent or by refusal – may, of course, result in biases of the samples. The data of this study will be archived in the Zentral Archive (Cologne) and the Steinmetz Archive (Amsterdam), and will become available to interested analysts by the beginning of 2001. Before that time, they can only be obtained with permission of the primary investigators.

Mannheim), Mark Franklin (Trinity College, Hartford, Conn., USA), Michael Marsh (Trinity College, Dublin, Ireland), Renato Mannheimer (University of Genoa), Jacques Thomassen (University of Twente, the Netherlands), and Wouter van der Brug, Holli Semetko, Klaus Schönbach and Cees van der Eijk (all from the University of Amsterdam). Relevant contextual knowledge and assistence with translations of questionnaires was provided by a series of ad hoc country specialists and native speakers, whose contribution has been of great value. The fieldwork was conducted by an international consortium of survey institutions, coordinated by IPSOS (Hamburg).

The funding for this large-scale study came from different sources. The major part was financed by the University of Amsterdam, supplemented by a large grant from NWO (the Netherlands’ Science Foundation). Important further contributions came from CIS (Madrid), the University of Genoa, the University of Mannheim, and Trinity College (Hartford, Conn., USA).

The Swedish sample in this study resulted in 505 successful interviews, that provide the data to be analyzed below.

The structure of the analyses

From the interview data, measures can be derived that pertain to the different approaches for explaining voter participation and that were described in Part 1 of this report. For each of these I will first of all report the differential in turnout between its categories, as these differentials manifested themselves in the responses of the Swedish voters in our study.19 This differential is simply the dif-


All analyses reported here were conducted on unweighted data, that means that no corrections were made to compensate for any kind of distributional differences between the sample on the one hand and the sampled population of Swedish adults on the other hand. Whereas weighting often affects univariate distributions, it has in general a much smaller impact on bivariate results (such as differences between groups) or multivariate analyses.

ference in percentage points between the first and the second mentioned categories of each measure. It describes differences in turnout, which are very helpful in indicating which of these characteristics of voters matter more, or less, in terms of differences in electoral participation. Yet, these differentials cannot without further ado be (causally) attributed to the measures in question, because of the possibility that the differences between the categories reported may be generated by other important characteristics, that may more plausibly be thought of as causes. Therefore, additional analyses will be presented, in which several measures are used simultaneously to explain (in statistical parlance) the difference between turning out to vote and abstaining. These multivariate analyses will first of all be presented for each of the clusters of measures that relate to the theoretical approaches discussed earlier, and subsequently for all these jointly.

Differences in turnout

The data of the EES99 contain information about a large number of characteristics of voters. As a start of the analyses to be reported, Tables 1 through 3 present differences in turning out to vote between different groups of respondents. Table 1 reports such differences in when looking at background characteristics of citizens.

Table 1: Differences in electoral participation by background characteristics, European Elections of 1999

(Data: Swedish segment of EES99, n=505)

Background characteristics

Groups compared Turnout differential (%)

Gender (men vs. women) 0.5 (n.s.) Education (high vs. low)) 4.8 (n.s.) Age (30 to 60 vs. under 30) 13.5 Age (over 60 vs. 30 to 60) 26.4 Union member (yes vs. no) -7.1 (n.s.) Attends church (never vs. rest) -13.6 Urbanisation (large town vs. rural area) 16.3

Table 1 shows that gender and education matter very little in terms of turnout. The difference in turnout between men and women is a mere 0.5 %, a difference that is statistically not significant, as indicated by the ”n.s.” designation in the last column.20 In other words, the difference is so small that it could have occurred by chance, without any real difference existing in the population. The difference in turnout for the two education groups is not significant either, which seems at first surprising. Two things must be kept in mind, however. First, compulsory education far into the teenage years ensures that the very large majority of the population is endowed with the cognitive and information-handling skills that education is expected to generate, with the consequence that


Unless indicated otherwise, in all analyses a p-level of .05 has been used to establish whether or not results are statistically significant. It must be kept in mind that whether or not a difference is significant depends both on the magnitude of the difference and the size of the groups that are compared. The latter are constrained by the total sample size, which is 505. The implication is that differences that are not significant in this study, may be found to be significant in other studies with a substantially larger number of respondents.

differences in this respect are smaller than when large portions of a population have barely any formal education at all. Second, education is measured rather crudely in this study, which also may mask differences that – in spite of the generally high level; of education – could possibly still be found with more refined measurements.

21The other background variables reported here are, however, very powerful in discriminating between those who voted and those who abstained. People over 60 years of age vote more frequently than those between 30 and 60 (a difference of 26.4 %), while the latter show a 13.5 higher turnout rate than those under 30. Combining these two differences shows that the oldest of these age groups has an almost 40 % higher turnout rate than the youngest!

People who sometimes attend church (irrespective of the frequency thereof) are considerably more likely to vote than those who say they never to go to church, a finding that seems in line with the social integration hypothesis mentioned in Part 1. This, however, is at odds with the fact that union-members are less, rather than more likely to vote (although this latter difference is not significant, possibly because of the relatively small group of members).

Finally, of this group of variables, is urbanisation. Respondents themselves indicated whether they lived in a large town, a small town or a rural area. The difference between the first and the last is significant (16.3 %), with the respondents living in small towns falling about halfway between the other two groups. Exactly why these differences between more and less urbanised respondents exist is not clear from these descriptive results; but it is clear that the differences are quite pronounced.

21 Education is measured by the age at which the respondent stopped with

full-time education. This way of measuring obviously does not represent kinds of education or levels of scholastic achievement. In the analyses here, the responses were divided into two categories: those that ended full-time education before the age of 18 (the ”low” group), and those that ended their education at age 18 or later (the ”high” group).

The extent to which a second group of respondent characteristics is correlated to turning out to vote is reported in Table 2. These variables all concern aspects of interest in and information and knowledge about politics. These variables form part of the syndrome of civic attitudes and orientations referred to in part 1. The variables that relate to media exposure are also of relevance in the approach that explains turnout by the efforts of mobilizing agents.

Table 2: Differences in electoral participation by political involvement and knowledge, European Elections of 1999

(Data: Swedish segment of EES99, n=505)

Political interest and knowledge

Groups compared Turnout differential (%)

Interest in the EE campaign (very/some vs. little/no) 32.0 Political interest / general (very/some vs. little/no) 32.3 Pays attention to political news (a lot/some vs. little/no) 35.9 Watched tv programs about EE

(often/sometimes vs. never) 30.4

Read papers about EE (often/sometimes vs. never) 31.6 Watches news on TV / general (always vs. =< 3 days p.w.) 31.0 Reads newspapers / general (always vs. =< 3 days p.w.) 21.1 Knows which party won EE (yes vs. no) 28.8 Mentions issue (yes vs. no) 14.7 Mentions party (yes vs. no) 9.3 Informed about EU (sufficient vs. insufficient) 6.1 (n.s.)

Table 2 shows that people who are interested in the European elections and its campaign, who are interested in politics in general, who keep informed by watching TV-news and reading newspapers about this specific election, or about politics in general, turn out in much higher numbers than those who are not interested, who pay no attention to political news, and who are exposed to little information from televison or newspapers. Almost all measures of interest and exposure – irrespective of whether they

are focused on the European election in particular – yield a turnout differential of approximately 30 %, which suggests that these different characteristics overlap to a considerable extent. Separate analyses (not reported here) show that this is indeed to a large extent the case: the distinction in the sample effectuated by one of the variables (say, interest in the EE campaign) coincides to a very large extent with the distinction that is made by another (such as political interest in general).

Three of the variables relate to knowledge and insight. The first is straightforward: whether or not people were able to indicate which of the Swedish parties had won the European election (only 45 % were able to correctly indicate this). The next two variables are ”proxies” for knowledge and insight. They indicate whether or not respondents were able to give any substantive answer to an open question about the most important problems in the country, respectively, which of the parties in the country would be most likely to undertake a preferred policy on that issue. The logic of these variables is that in order to provide any kind of answer, respondents need some idea (which may be entirely idiosyncratic) of what goes on in the country and of what the various parties stand for. Both variables are significantly related to turnout, but the differential is much smaller than in the case of the interest and exposure variables.

The last variable in this table concerns whether the respondents feel that they are sufficiently or insufficiently informed about the politics of the European Union. Although those who think that they are sufficiently informed (less than 20 % of the sample) turn out in higher numbers, the difference is not statistically significant.

A final series of variables, reported in Table 3, deals with political preferences, civic attitudes and political approval and satisfaction.

Political preferences, first of all, relate to the extent that respondents identify with a political party, or the extent to which they find one of the parties electorally appealing. Those who identify with a party turn out 23.7 % more often than those who do not. Moreover, the stronger one’s preference for a party – irres-

pective which – the larger the likelihood that one turns out to vote. For each of the parties in Sweden, the respondents were asked how likely it was (on a scale from 1 to 10) that they would ever vote for that party. For each person it is thus possible to derive from his/her entire set of responses which party he/she finds the best one of all, and how strong that preference is (measured on the same scale of 1 to 10). The stronger this preference, the more likely it is that the respondent turned out to vote: if the ”best” party scores a 10, the probability of voting is more than 12 % higher than if the best party scores a 9, etc. In other words, the ”better” the best party is in the voter’s eyes, the more likely he/she is to vote. As this has little to do with the specific context of the election, or with the efforts of mobilizing agents in that election – as is equally the case for identifying with parties – this shows that the stronger the affective relation between voters and parties the larger the likelihood of voting.

Intention to vote in a national election is also strongly linked to turning out to vote in the European election. Together with the previous findings in this table this demonstrates that turnout in European elections is very much related to a voter’s affective linkage to the national party system.

Table 3: Differences in electoral participation by political preferences, civic attitudes, political approval and satisfaction, European Elections of 1999

(Data: Swedish segment of EES99, n=505)

Groups compared

Turnout differential (%)

Political preferences

Identifies with a political party (yes vs. no)


Preference for ”best” party (10 vs. 9)


Preference for ”best” party (10 vs. 8)


Preference for ”best” party (10 vs. 7)


Votes in national election (yes vs. no)


Civic attitudes

Politics too complicated (agree vs. rest) -22.7 So many voters (agree vs. rest) -26.7 Parties not different (agree vs. rest) -11.6

Political approval and satisfaction Approve of national government (approve vs. rest) 7.8 (n.s.) Satisfaction with democracy in Sweden (satisfied vs. rest) 27.1 Satisfaction with democracy in the EU (satisfied vs. rest) 15.5 Satisfaction with policy on EU integration

(satisfied vs. rest) 14.6

Opinion on EU integration (further vs. too far) 31.3 EU is good/bad thing (good vs. bad) 31.2

The three items referred to as civic attitudes refer to three statements of which the respondents had to indicate whether or not he/she agrees with them. These three are, in order:

Sometimes politics is so complicated that someone like me just cannot understand what is going on.

So many people vote in elections that my vote does not matter.

Most of the parties in Sweden are so much alike that it does not make much of a difference which one is in government.

In different ways, agreeing with each of these statements expresses a feeling that it does not make sense to take part in elections. No wonder that in all three cases those who agree with the statement vote in significantly smaller numbers than others.

Political approval and satisfaction, as long as they are of a general kind, have a strong relation to turnout. As soon as the satisfaction or approval becomes tied to political parties – as in the case of approval of the government or satisfaction with (government) policy on European integration – it is less or not significantly related to voting versus abstaining. Those not satisfied with democracy in Sweden are considerably less likely to vote than those who are satisfied. This differential is much smaller for the question about satisfaction with democracy in the EU, which demonstrates that not being satisfied with democracy in Europe is less a deterrent to voting than not being satisfied with democracy in Sweden itself. This makes sense: after all, one is represented in the European Parliament by way of the Swedish parties, and as long as one is content with those, one may as well vote.

Disapproval of the EU and of European integration is, however, very much a stimulus to stay at home in the European elections. Those who find the EU a ”bad” thing (approximately one third of the sample), or who find that integration has already gone ”too far”, have more than a 30 % lower turnout rate than citizens who regard the EU as a ”good” thing (also about one third of the sample) and those who feel that integration should be ”pushed further”. This suggests that, to a considerably extent, the European elections in Sweden are elections for a system that a large segment of the population disagrees with. Representation via the European Parliament is evidently not seen by this segment as a way to improve matters, and, in view of the pro-integrative stance of the EP, not entirely without reason.

Multivariate Analyses

The analyses reported in Tables 1, 2 and 3 are suggestive, but limited in the conclusions that can be drawn from them. Two problems in particular play a role. First, there may be a considerable

”overlap” between the various characteristics that were investigated. Second, even though the turnout differential between groups compared may be large, this difference is not necessarily caused by the characteristic on which the groups differ. In order to deal with these problems, multivariate analysis is necessary; that is, analysis in which a number of different characteristics are simultaneously taken into account in terms of their consequences for turnout. In addition to this, theoretical considerations must be utilized, on the basis of which various factors can be ordered vis-a-vis each other as more distant, or intermediary causes of turnout. I will report here only a limited number of such analyses, using multiple regression methods.22

The investigations consisted of a series of consecutive analyses, in which voter characteristics that were shown to be ”overlapping” in their effect on turnout (in statistical parlance: factors that are multicollinear) were weeded out to such an extent that only the most important of them are retained, and the superfluous other ones eliminated. To the extent that – from a statistical point of view – multicollinear variables are equally important, I chose to retain the most generic one and to drop the more specific one.23These analyses were done first for each of the blocks of variables distinguished in Tables 1 to 3 (1 block of variables each from Tables 1 and 2, 3 from Table 3). The variables that remained from these analyses (all of them significant within these separate analy-

22 An objection against linear methods such as multiple regression is that

this technique is not optimally suited for a dichotomous dependent variable, as turnout is. More appropriate methods, of the non-linear regression type, are, however, found to rarely differ in substantive results from the method used here (e.g., Oppenhuis, 1995, Franklin, Van der Eijk and Oppenhuis 1996). Moreover, they are much less self-evident for a non-technical audience than traditional regression methods. Explicit causal models, in which an entire structure of hypothesized causal relationships is tested, would be the logical next step following the analyses reported here. Such more elaborate analyses will be reported elsewhere, however.


For a more detailed description and justification of the strategy of analysis, see Franklin, Van der Eijk and Oppenhuis (1996).

ses) were then successively combined into a final model, the result of which is a multivariate regression model that is reported in Table 4, and will be discussed below.24

Table 4 shows in what sequence the variables of the five blocks were entered and, for each block or addition, the explanatory power of the model (see the column of cumulative adjusted R2). All together, these variables explain 29 percent of the variance, and each of the blocks adds significantly to the explanation, over and above what had already been explained by variables previously entered. Although 29 percent of variance explained in turnout-differences does not seem very impressive at first sight, it is actually not a poor result in comparison to similar analyses reported elsewhere.25It does indicate, however, that individual electoral participation is only to a limited extent explicable by the kind of factors discussed; a more complete explanation would in all likelihood require many rather idiosyncratic factors. On the other hand, the results also imply that a small number of voter characteristics accounts to a large extent for turnout differences in a systematic manner.

24 These exploratory analyses were run with pairwise deletion of missing

data; listwise deletion yields virtually identical results.


See, e.g., van der Eijk and Oppenhuis (1990), who in an somewhat similar analysis for the Netherlands and the European election of 1989, get to an R


of 25 percent. In an EU-wide analysis for 1989 Oppenhuis

(1995) arrives at an R


of slightly over 20 percent. Moreover, it would be

fallacious to assume that 100 percent is the attainable maximum in explained variance. For a number of technical reasons, relating to skewed distributions of categorical data, the attainable maximum is much lower, often in the range of 50 to 60 percent, but difficult to calculate with precision.

Table 4: Multivariate regression of electoral participation in the European Elections of 1999

(Data: Swedish segment of EES99, n=505)

beta significance Cumul. adj. R2

Block 1





.06 n.s. (p=.16)

Attends church

.02 n.s. (p=.60)

Block 2


EU is good thing


Satisfied with democracy in Sweden .12

Block 3


Identifies with a political party

.07 n.s. (p=.08)

Preference for ”best” party


Block 4


Interest in EE campaign


Pays attention to political news .11 Reads papers about EE .09

Block 5


Politics too complicated


So many voters


The column with beta coefficients indicates the relative explanatory strength of the different variables. The higher this coefficient, the more important the variable in question in explaining why some people vote and others do not. In some instances, the multivariate perspective of this table results in variables losing their significance compared to what was seen in Tables 1–3. This does not necessarily mean that such variables are not important in the explanation, but rather that their effect on turnout runs via one of the other variables that is now included in the same analysis. This is the reason why, for example, church attendance and urbanization are not significant in Table 4. The effect of the urban-rural distinction on turnout is certainly real, but it is largely mediated by opinions about the EU (in urban areas people are more inclined to think the EU is a ”good” thing, in rural areas people tend much

more to seeing it as a ”bad” thing), and by differences in political interest and involvement (in rural areas party identification is lower, people pay comparatively less attention to politics, and agree more often with the statement that politics is too complicated to understand). Likewise, the effect of church attendence is taken over by satisfaction with democracy in Sweden, by interest in the campaign for the European elections and by finding the EU a ”good” thing.26

Table 4 leads to the following observations. First, a clear effect exists of opinions about the EU (”good thing” vs. ” bad thing”), even when the other explanatory factors mentioned in this table are taken into account. Such an effect was not found in any of the other members of the European Union – at least as far as the 1989 and 1994 European elections were concerned.27 The effect may be not surprising in the light of the controversies in Sweden over membership of the EU and related matters (such as the common currency), but it raises the question whether in 1999 this is a specific Swedish phenomenon, or a pattern that has emerged elsewhere as well. This report is not the place to investigate this question, and analyses about this will be reported elsewhere.

Not only evaluative opinions about the EU have an effect on electoral participation, but also evaluations of Swedish democracy itself. Almost 25 % of the sample indicates that they are not very, or not at all satisfied with democracy in Sweden, and this opinion


Interestingly, a strong correlation exists between being satisfied with democracy in Sweden and seeing the EU as a ”good” thing. This suggests that a divide exists in the Swedish population, with on the one extreme a segment that does not approve of the EU and is dissatisfied with democracy in Sweden (most strongly reprresented in the countryside) and on the other extreme a segment that is satisfied with democracy in Sweden as well as with its membership of the EU (most strongly represented in the urban areas). As far as evaluations concerning the EU is concerned, this is not a new insight. That it is linked to (dis)satisfaction with democracy in Sweden is less trivial, as well as politically more disturbing. This report, however, is not the place to elaborate these findings further.


On this point see Franklin, van der Eijk and Oppenhuis (1996) and Van der Eijk, Franklin and Mackie (1996, p. 275).

is not only related to staying at home in the European elections of 1999, but also to feeling that Swedish membership of the EU is a ”bad” thing.

Being integrated into the social and political system is also clearly related to voting versus abstaining, quite in accordance with theoretical expectations described in part 1 of this report. Social integration is to some extent captured by church attendance, and being politically integrated by party identification, but the latter is captured even better by the question of whether the system contains at least one party that a respondent feels really positively about (the variable referred to in Table 4 as ”Preference for ’best’ party”).

Interest in the European election campaign, reading about this election in newspapers, and paying attention in general to political news, are also strong determinants of turnout. Surprisingly, in the light of findings for other EU members in previous European elections, the specific items about interest in the EE campaign are more powerful than general indicators of political interest. This suggests that, in contrast to other countries in 1989 and 1994, the European election in Sweden in 1999 contained an element that cannot be entirely reduced to the voter’s links to the domestic political system. Again it will have to be seen in further research whether this is also to be found in other countries in 1999, and, again, such analyses will have to be reported elsewhere.

Finally, in accordance with the theoretical expectations expressed in Part 1, feelings of political (in)efficacy do explain some of the differences between voting and non-voting, even when all other factors in Table 4 are simultaneously taken into account.

Part 3: Reflective comments and informed speculation

The empirical analyses contained in this brief report are limited. This is for a number of different reasons:

Information about the mobilising efforts (or lack thereof) by political parties, media and other politically relevant actors, organizations and institutions is at this moment not available. Therefore, it is impossible to assess the kinds of forces that were referred to in part 1 as mobilising agents. This is not to say that these considerations are irrelevant when trying to explain the low turnout in the 1999 European elections in Sweden, on the contrary. In Sweden, as in the other member states of the Union, parties, political leaders, relevant interest groups and media displayed a distinct lack of interest and activity, particularly when compared to their behavior at the time of elections for the Riksdag. This is, obviously, not the entire explanation for the low turnout in the European elections, but it should nevertheless not be ignored as a relevant factor. The drop in turnout between the first (1995) and the second (1999) European elections in Sweden may possibly be related to a difference in the level of activities of these political and social institutions. All member states of the European Union experienced a distinctly higher turnout in their first than in their second and later European elections. This first-time boost seems at least to some extent attributable to the extra attention that is given to the novel phenomenon of European elections, a factor that is absent in all subsequent elections to the European Parliament.

A fuller explanation for the low level of turnout in the European election in Sweden in 1999 requires comparative analyses, where the comparisons involve both other elections in Sweden (such as previous European elections, various Riksdag elections, etc.), and European elections in other countries (in 1999, as well as at previous European elections. Obviously, this is not the place for such extensive comparisons, especially since much of the information that would be required is not, or not yet,

available. The approach, referred to in Part 1, that seeks to relate levels of turnout (as well as individual differences in electoral participation) to ”what is at stake” in various elections, is one that can only be employed in such a comparative approach – just like the approach that focusses on mobilizing agents, discussed in the previous paragraph. Because of these limitations, the results of the empirical analyses reported here pertain only to individual differences in behavior within the single context of June 1999 in Sweden. Every aspect of that context is thus a constant in this study and, even though these aspects may have contributed to the low over-all turnout (in Swedish perspective), they remain inaccessible for research as long as they are constants.

As indicated several times in Part 2, more elaborate analyses can be performed on the available Swedish data. Such analyses will explicate in greater detail the inter-relationships between the factors that influence individual electoral participation, and will thus help us to understand more precisely how factors such as, e.g., age, are related to differences in turnout. They will complement, but not replace, the major empirical findings reported in Part 2.

Apart from pointing out the limitations of the analyses in this report, a bit of informed speculation seems in order to shed tentative light on the declining level of turnout in Swedish elections, for the

Riksdag as well as for the European Parliament. The remarks below

are to some extent based on earlier research28 and to some extent on logical extrapolation from the interpretations of that earlier research.

Instrumental motivations are of great importance in explaining turnout in contemporary western electorates. From this follows the logic that turnout will be lower when the less is at stake in an


The most important previous research that I use here is Van der Eijk and Franklin (1996), Oppenhuis (1995), Franklin, Van der Eijk and Oppenhuis (1995), Schmitt and Thomassen (1999), Franklin (1996), and some of the sources referred to in these texts.

election. That there is little at stake in European elections is not because of the alleged limitations of the powers and prerogatives of the European Parliament. After all, the role of the EP has grown substantially in recent years, most recently since the ratification of the Treaty of Amsterdam. The European Parliament has currently a far from negligeable role in the legislative process of the EU. This is evidenced by the fact that those who have a specific interest in certain EU legislation – such as branches of industry, social movements, and subnational government agencies – are willing to invest considerable resources trying to influence the results of EU legislation via lobbying at the EP. Yet, despite this increased role of the EP in the making of EU legislation, the power of the EP does not extend to the composition of the European Commission, let alone to that of the Council. Stated differently, European elections decide the composition of the European Parliament, which in itself is important for the outcome of individual pieces of legislation, but which is not in any way of relevance for the composition of the executive and policy-initiating institutions of the EU. In contrast to the situation that exists in the elections for the national parliaments of the EU member-states, it is absolutely impossible to link the composition of the EP to any kind of over-all policy direction that will be in force for the subsequent period. Moreover, the absence of a government vs. opposition divide in the EP makes it impossible for ordinary citizens (as well as for most media) to link the results of the elections for the EP to any kind of policy direction of the EU. These circumstances thwart almost any sort of instrumental motivation for casting a vote in the European elections.

One could wonder why, given the logic described above, turnout in European elections was generally higher in previous European elections than in 1999.29 Apart from the drop in turnout after the boost that occurs only at the first time of European elections, turnout has continued to decline in most member-states, albeit


The arguments hereafter originate from intensive discussions on these topics between Mark Franklin, Michael Marsh and myself, conducted at various occasions.

with occasional increases, that often were attributable to occasional concurrency of the European elections with other elections in one of the member-states, as was the case with the Irish Dàil elections in 1989, and with regional elections in Spain in 1999. Two different mechanisms can be expected to operate here. One is generational replacement, by which former generations – socialized in conditions of cleavage politics and the resulting high degrees of electoral mobilization – are gradually replaced by new generations that were never subjected to such intense mobilizing efforts, and that have therefore a weaker internalization of the norm that one ”should” vote. The second mechanism that would lead to the persistent decline in turnout for European elections may be thought of as gradual learning. It takes time before the default interpretation of political systems along the patterns that one is most used to (i.e., the system of one’s country) is experienced as inadequate when dealing with the European Union.

A final remark that has to be made here concerns turnout in elections for the national parliaments of the member states of the EU. Although Topf described as late as 1995 the variations in turnout in these and other western countries as ”trendless fluctuation”, more recent observations of turnout in national elections in EU countries point in the direction of a gradual decline. If this is indeed the case, one could, of course, point here too to the still further declining remnants of cleavage politics and generational replacement. In addition to this, however, one could think of a more structural reason that, if this speculation is correct, will have additional depressing effects on turnout in national parliamentary elections. If citizens learn as a consequence of being exposed to reality, they not only learn that the results of elections for the EP are unrelated to any kind of policy direction for the EU, but they also learn that their own governments (and the parliamentary majorities upon which these rest) become increasingly less influential for a growing number of policy areas, as EU legislation and regulation become more important and limit the scope of national policy-making. As the EU syphons off policy making powers from national parliamentary regimes, the stakes in national elections become smaller, and the same instrumental logic that generates

declining turnout in EP elections will equally give rise to declining turnout in national elections.

Such a development would not be detrimental from a normative democratic point of view if the diminished importance of the (party) composition of national parliaments would be compensated by an increased importance of the party composition of the European Parliament. As argued above, however, this is not at all the case because of the current institutional structure of the EU. Seen from this perspective, the citizens of the EU are gradually disempowered electorally, and the declining turnout in national and European elections can only be regarded as a rational reaction of the voters to this development. In another context I argued with my co-authors that the conditions that turn European elections into so-called second-order elections undermines not only the democratic character of the EU, but tends to do so for domestic political processes as well (Franklin, van der Eijk and Marsh, 1996). A declining impact of elections on policy-direction and a declining participation of citizens in the electoral process – at the national as well as the the EU level – can only increase this danger.


Biorcio, R. and R. Mannheimer (1995) Relationships between

Citizens and Political Parties. In H.-D. Klingemann and D. Fuchs, eds., Citizens and the State. Oxford: Oxford University Press; pp. 206–226 Blumler, J.G., ed. (1983) Communicating to Voters. Popular

Responses to the First European Community Elections. London:

Sage Brug, W. van der (1997) Where’s the Party? Voters’ Perceptions of

Party Positions. Dissertation, University of Amsterdam

Dalton. R.J. and M.P. Wattenberg, (1993) The Not so Simple Act

of Voting. In A.W. Finifter, ed., Political Science: The State of

the Discipline II. Washington: American Political Science Asso-

ciation Egmond, M. van (1999) Rain Falls on All of Us (but Some Manage

to get more Wet than Others). Accounting for Context Through the Individual in Electoral Participation. Paper, presented at

EPOP, Northampton, September 1999 Eijk, C. van der, and W. van der Brug (1998) Different at the Sur-

face, Similar in Depth – Attitudes towards Politics in Germany and the Netherlands, Amsterdam: DIA

Eijk, C. van der, and M.N. Franklin, eds. (1996) Choosing Europe?

The European Electorate and National Politics in the Face of Union. Ann Arbor (Mich.): The University of Michigan Press

Eijk, C. van der, and E.V. Oppenhuis (1990) Turnout and Second-

order Effects in the European Elections of June 1989 – Evidence from the Netherlands. Acta Politica, vol. 25, nr.1:67– 94 Eijk, C. van der, M.N. Franklin and T. Mackie (1996) The Dog

that Did Not Bark: The 1994 Elections in the Light of 1989. In C. van der Eijk and M.N. Franklin, eds., Choosing Europe? The

European Electorate and National Politics in the Face of Union.

Ann Arbor (Mich.): The University of Michigan Press; pp. 268–282

Franklin, M.N. (1996) Electoral Participation. In L. LeDuc, R.G.

Niemi and P. Norris, eds. Comparing Democracies. Elections

and Voting in Global Perspective. Thousand Oaks (Cal.): Sage;

pp. 216–235 Franklin, M.N., C. van der Eijk and M. Marsh (1996) Conclusions:

The Electoral Connection and the Democratic Deficit. In C. van der Eijk and M.N. Franklin, eds., Choosing Europe? The

European Electorate and National Politics in the Face of Union.

Ann Arbor (Mich.): The University of Michigan Press; pp. 366–388 Franklin, M.N., C. van der Eijk and E.V. Oppenhuis (1995) The

Motivational Basis of Electoral participation: European Elections provide a Critical Test. Paper, presented at ECPR Joint

Sessions, Bordeaux Franklin, M.N., C. van der Eijk and E.V. Oppenhuis (1996) The

Institutional Context: Turnout. In: C. van der Eijk and M.N. Franklin, eds., Choosing Europe? The European Electorate and

National Politics in the Face of Union. Ann Arbor (Mich.): The

University of Michigan Press; pp. 306–331 Franklin, M.N., T. Mackie and H. Valen, eds. (1992) Electoral

Change. Responses to Evolving Social and Attitudinal Structures in Western Countries. Cambridge: Cambridge University Press

Granberg, D. and S. Holmberg (1988) The Political System Matters:

Social Psychology and Voting Behavior in Sweden and the United States. Cambridge: Cambridge University Press

Gunsteren, H. van, and R.A. Andeweg (1994) Het grote Ongenoe-

gen: over de Kloof tussen Burgers en Politiek. Haarlem: Aramith

Kaase, M. and K. Newton, eds. (1995) Beliefs in Government.

Oxford: Oxford Univeristy Press Lijphart, A. (1968) The Politics of Accommodation: Pluralism and

Democracy in the Netherlands. Berkeley: University of Califor-

nia Press Lipset, S. M., and S. Rokkan (1967) Cleavage Structures, Party

Systems and Voter Alignments. In S. M. Lipset and S. Rokkan,

eds., Party Systems and Voter Alignments: Cross-National

Perspectives. New York: Free Press

Marsh, M., and M.N. Franklin (1996) The Foundations:

Unanswered Questions from the Study of European Elections, 1979–1994. In C. van der Eijk and M.N. Franklin, eds.,

Choosing Europe? The European Electorate and National Politics in the Face of Union. Ann Arbor (Mich.): The University of

Michigan Press; pp. 11–32 Mill, J.S. (1974/75) A System of Logic. Vols. vii and viii in

Collected Works of J.S. Mill. London: Routledge and Kegan Paul

(originally published in 1843) Oppenhuis, E.V. (1995) Voting Behavior in Europe. Amsterdam:

Het Spinhuis Parry, G., G. Moyser and N. Day (1992) Political Participation and

Democracy in Britain. New York: Cambridge University Press

Przeworski, A., and H. Teune (1970) The Logic of Comparative

Social Inquiry. New York: Wiley

Reif, K., ed. (1985) Ten European Elections: Campaigns and Results

of the 1979/81 First Direct Elections to the European Parliament.

Aldershot: Gower Reif, K., and H. Schmitt (1980) Nine Second-Order National

Elections. A Conceptual Framework for the Analysis of European Election Results. European Journal of Political Research, 8:3–44 Rose, R., and D. Urwin (1970) Persistence and Change in Western

Party Systems since 1945. Political Studies, vol. 18:287–319 Schmitt, H., and J. Thomassen, eds. (1999) Political Representation

and legitimacy in the European Union. Oxford: Oxford Univer-

sity Press Siegfried, A. (1913) Tableau Politique de la France de l’Ouest sous

la Troisième République. Paris: Colin

Tingsten, H. (1937) Political Behavior: Studies in Election Statistics.

London: King and Son

Topf, R. (1995) Electoral Participation. In H.-D. Klingemann and

D. Fuchs, eds., Citizens and the State. Oxford: Oxford University Press; pp. 27–51 Verba, S. and N. Nie (1972) Participation in America. Political

Democracy and Social Equality. New York: Harper and Row

Verba, S., N. Nie and J.O. Kim (1978) Participation and Political

Equality. Cambridge: Cambridge University Press


European Parliament Elections

Institutions, Attitudes and Participation

Richard Sinnott1


This paper begins with a description of the problem of low participation in European Parliament elections and, in the course of these observations, it makes some preliminary suggestions as to the causes of variation in turnout. In order to place the problem of low turnout in context, the paper then looks briefly at attitudes to European integration and at the relationship between the citizens and the Union. The issue of the discrepancies in prevailing interpretations of the sources of low turnout in European Parliament elections is then considered. A concluding section tackles the question of what should be done about the problem of abstention, considering the matter under two headings: the facilitation of participation and the mobilisation of participation.

The variations in turnout in European Parliament elections

Over the period since 1979 and as a direct result of successive revisions of the EU treaties, the powers of the European Parliament

1 The author is professor of Politics at Department and Centre for

Comparative Research on Public Opinion and Political Behaviour at University college Dublin.

have increased significantly. In contrast to the increasing power and importance of the Parliament, turnout in European Parliament elections has declined – from an average turnout of 65.9 per cent in 1979 to 52.4 per cent in 1999. In fact the picture is worse than these figures indicate. Three countries in the European Union (Belgium, Greece, and Luxembourg) have compulsory voting and Italy had compulsory or at least quasi-compulsory voting until 1993. Furthermore certain countries either regularly or from time to time hold concurrent nation-wide elections (national, regional or local) that artificially boost the turnout in the European election. Leaving these countries to one side, turnout in the remaining member states was 52.9 per cent in 1979 and only 39.4 per cent in 1999 (see Table 1).

Apart from the boost arising from compulsory voting and the concurrence of other elections, turnout in European Parliament elections is affected by the day of voting (Sunday versus weekday), with weekday voting contributing to low turnout in European Parliament elections in Denmark, the Netherlands, Britain and Ireland. Beyond these few obvious factors , however, there is widespread disagreement as to what causes low turnout in European Parliament elections. This issue will be analysed in some detail below. But first it is necessary to consider the matter of overall attitudes to European integration among the citizens of the Union.

Table 1. Turnout in five European Parliament elections, 1979–99


European elections

Mean turnout

1979 1984 1989 1994 1999 1979-99


91.4 92.2 90.7 90.7 90.0 91.0

Luxembourg 88.9* 88.8* 87.4* 88.5* 85.8* 87.9 Italy 84.9 83.4 81.0 74.8 70.8 79.0 Greece n.a. 77.2 79.9* 71.7 70.2 74.8 Spain n.a. n.a. 54.6 59.1 64.4* 59.4 Germany 65.7 56.8 62.3 60.0 45.2 58.0 Ireland 63.6* 47.6 68.3* 44.0 51.0* 54.9 France 60.7 56.7 48.7 52.7 47.0 53.2 Denmark 47.8 52.3 46.2 52.9 50.4 49.9 Austria n.a. n.a. n.a. n.a. 49.0 49.0 Netherlands 57.8 50.6 47.2 35.6 29.9 44.2 Portugal n.a. n.a. 51.2 35.5 40.4 42.4 Sweden n.a. n.a. n.a. n.a. 38.3 38.3 U.K. 32.3 32.6 36.2 36.4 24.0 32.3 Finland n.a. n.a. n.a. n.a. 30.1 30.1

Mean – all member states

65.9 63.8 62.8 58.5 52.4 56.3

Mean – states without compulsory voting or concurrent nation-wide elections

52.9 49.4 49.5 47.0 39.4 47.6

*Concurrent nation-wide elections

Attitudes to European integration

The measurement of support for European integration is highly sensitive to variations in question wording. The four standard Eurobarometer indicators of attitudes to integration are customarily labelled ”unification”, ”membership”, ”benefits”, and ”dissolution”. The levels of support for integration elicited by these questions depend both on the stimulus presented in the question and on the response categories, in particular on whether the response categories offer an explicit middle position.

The unification indicator measures support for a very general aspiration (”efforts to unify Western Europe”) and does so on a four point scale (”very much for”, ”to some extent for”, ”to some extent against”, ”very much against”) that does not provide an explicit middle or neutral point. The rather vague stimulus and the absence of a middle position combine to produce high levels of support for integration, touching almost 80 per cent at the beginning of the 1990s (see Figure 1). The membership indicator provides a more concrete stimulus (country X’s membership of the European Union) and a three point scale that includes an explicit middle position (”a good thing”, ”neither good nor bad”, ”a bad thing”). As such, it probably provides a more realistic gauge of support for integration that typically runs some 10 to 20 percentage points behind the unification indicator. The third Eurobarometer indicator asks whether the respondent’s country has benefi-

ted from membership of the Union but, like the unification indi-

cator, it does not provide a middle position (the response categories are ”benefited” and ”did not benefit”). Because it involves an element of perception as well as evaluation, it is not surprising that the benefits indicator registers a positive response that is slightly lower than that registered by the membership indicator. Indeed, the gap between the two might be greater were it not for the presence of a neutral category in the membership indicator and the absence of such a category in the benefits indicator. Finally, there is the dissolution indicator. This question poses the hypothetical situation of the scrapping of the Community or Union, with strong negative and positive options and a middle position (”very sorry”,

”indifferent”, ”very relieved2”). The indicator has been criticised because it is hypothetical; nonetheless it provides a useful measure of enthusiasm or lack of enthusiasm for European integration, showing, in June–July 1994 for example, a quite modest level of enthusiasm (43 per cent) that was actually slightly lower than the level of indifference and don’t know combined (46 per cent) (see Figure 1).

Figure 1. Attitudes to the European Union four Eurobarometer indicators (unification, membership, benefits and dissolution) 1973–1998

2 The ”very” was added to ”relieved” in 1993.

None of these questions is an ideal indicator of attitudes to integration; nor could it be said that they together form an adequate battery of items. One must therefore exercise a certain degree of caution in interpreting the evidence. In particular, attempts to categorise the indicators in terms of diffuse-affective versus specificutilitarian dimensions of support quickly run into difficulties. As Niedermayer (1995; 54-5) notes, the only unambiguously utilitarian measure is that based on the benefits question. That having been said, however, these are in many respects the best available data and, provided they are interpreted cautiously, they enable one to make some reasonably valid inferences. Before attempting a summary of the current state of support for European integration as evidenced by these indicators, one must also take account of the substantial variation in the indicators over time and across the member states.

In terms of changes over time, there was first of all a significant falling off in support for integration between the late 1970s and the early 1980s (see Figure 1). This may have been related to the prevailing Eurosclerosis that many commentators on European integration have identified. This was followed, however, by a substantial and sustained rise in support between 1982 and 1991. It is worth noting that this rise predated the arrival of Jacques Delors as President of the Commission and certainly predated the major initiative of the first Delors presidency, namely the launch of the Single Market ”1992” programme. On the other hand, there can be little doubt but that the rise and rise in support for integration was sustained by the activism of the Delors Commission, by the passage of the Single European Act and by the publicity and promotional efforts that surrounded Project 1992.

In fact, however, the actual arrival of the calendar year 1992 confirmed a general downward trend in support for integration (see Figure 1). The period since 1989 is so packed with political and economic developments that it is impossible to attribute this decline to any one factor3. Indications that the decline was under-

3 There is a substantial and technically highly sophisticated literature on the (mainly economic) determinants of support for integration.

way in a number of member states even prior to 1991 (Niedermayer 1995: 67) suggest that it was not simply a response to the signing of the Maastricht Treaty but may also have reflected a negative reaction to the growing intrusiveness of the Single Market programme on both the politics and economics of individual states. Another factor was the waning of the euphoria that surrounded the fall of the Berlin Wall and the growth in the realisation that event brought challenges and uncertainties regarding the future shape and role of the Community and was not simply the dawning of a new era of peace and prosperity for all. Whatever its causes, the decline in support for integration that began in the second half of 1991 has shown little or no sign of being reversed

While overall support for the Union over time has varied by some 20 percentage points, the range in support across countries has been as much as 50 percentage points; in 1997, for example, support, as measured by the membership indicator, went from 31 per cent in Sweden to 83 per cent in Ireland. Although any grouping of countries based on definite cut-off points in a distribution such as this is somewhat arbitrary, it is probably useful to think of a group of four countries with high levels of support for EU membership and a group of six with quite low levels. The former comprises Ireland, the Netherlands, Luxembourg and Italy. The composition of the latter is more surprising: predictably, it includes the three most recent entrants to the Union (Sweden, Austria and Finland) and the UK; the surprise is that it also includes Germany and Belgium but does not include Denmark. On the other hand, actual opposition to integration is as high in Denmark as in all but one of the low-support countries.

Eichenberg (1998) provides an excellent review and extension of this research. From the point of view of the present discussion, the most relevant findings are that ”the influence of objective economic circumstances on support for integration was less in the period following Maastricht than it had been before” (Eichenberg, 1998:12) and that the effects vary depending on the indicator of the dependent variable that is used.

When one probes beyond the very general sentiments captured by the four basic indicators, to, for example, attitudes to whether or not certain policy areas should be decided at the European or at the national level, one finds even greater discrepancies between member states and a somewhat different ordering. Thus in regard to the issue of the Europeanization of defence policy, some countries show a very high degree of support for Europeanization and others a very low degree. Ranged on the side of the Europeanization of defence are the Netherlands (76 per cent) in first place and, at a somewhat lower level, Belgium, Luxembourg, Italy and Germany. On the other hand, Sweden and Finland are, on this evidence, overwhelmingly opposed to the Europeanization of decisionmaking in the defence area, Finland being 8 per cent for and 90 per cent against. On this negative side also, though not as strongly, one finds Ireland, Greece, Denmark, Portugal and the United Kingdom. However, before drawing any large inferences about the future prospects for a common European defence policy from these data, one should note that the rate of don’t know responses to this item is remarkably, one might say suspiciously, low (an average of 4 per cent).

The suspicion mentioned in the previous paragraph leads to the hypothesis that some significant proportion of the responses on preferences regarding the Europeanization of defence and other policies actually reflect non-attitudes. A Eurobarometer-based research project on turnout in the European Parliament elections of 1994 provided the opportunity to test this hypothesis by using an exploratory question on the overall issue of policy attribution between the national and European levels. The new question, which was inserted in Eurobarometer 41.1, included the specific response category ”I haven’t really thought about it”. This response was chosen by 26 per cent of respondents, a proportion that, when combined with the 10 per cent who spontaneously offered a ”don’t know” response, yields more than one-third of the sample who acknowledge that they do not have any opinion on the basic issue of the appropriate scope of decision-making competence of the European and national authorities (Blondel, Sinnott and Svensson, 1998, 65–72)). Further evidence from the same study suggests that

even those who were willing to express an opinion, despite, to echo Converse (1964), a generous invitation to disavow any opinion when none was felt, may not have had a very explicit or well thought-out basis for that opinion. Those who did take a view on the question of the overall range of issues decided on by the European Union were asked ”When you say (insert response to previous question), is this a general feeling that you have about the European Union (European Community), or have you specific issues in mind?” This retrospective probe showed that only 17 per cent of the sample had specific issues in mind in responding to the original question. This reinforces the view that attitudes in this area may be less than well formed.

This expectation is further confirmed by Eurobarometer data from 1995 on perceptions of the allocation of decision making power between national governments and ”the European Union level” over a wide range of issues. Inter alia, the data show that 38 per cent of the European public believe that foreign policy is decided at EU level and precisely the same proportion believe that defence matters are decided at EU level (Eurobarometer, 1995: B65). On any reading of the common foreign and security policy as of 1995, these perceptions were wildly inaccurate. The proportion seeing foreign policy as being ”at least to some extent decided at the European Union level” should be much higher and the proportion seeing the same for defence should be much lower. Lest it be assumed that people are getting it wrong in relation to the common foreign and security policy simply because this is an inherently complex and remote area, one should also note that the proportion perceiving agricultural policy as being ”at least to some extent decided at the European Union level” is only 40 per cent.

The overall impression from this review of the evidence of various aspects of support for European integration can be summarised as follows: there is fairly widespread support for the rather vague notion of ”efforts to unify Western Europe”; as a result of a downturn since the second half of 1991, support for membership of the Union is running at only about 50 per cent and is matched by an almost equal level of indifference as to whether or not the Union continues to exist; finally, despite initial appearances to the

contrary, attitudes to the policy scope of the Union are neither well-formed nor well-informed. All of this suggests that the ”permissive consensus” (Lindberg and Scheingold, 1970), if it ever existed, was a rather fragile creature. One could go further and suggest that the term itself, so much bandied about, was actually misleading in that it glossed over significant flaws in the fabric of public opinion towards integration. Were it not for the prevalence of the permissive consensus assumption, there might not have been such surprise when, as the integration process began to make greater inroads on the economic and political life of the member states, support for integration began to wane, a waning that became manifest not only in opinion polls but also in referendums and in parliamentary debates in several countries.

These weaknesses in attitudes to integration are confirmed by the evidence relating to knowledge of the institutions of the Union. Space does not permit a detailed discussion of this aspect but the essentials can be briefly summarised. Knowledge of European affairs is low; prevailing attitudes to the Union are characterised by either indifference or lack of knowledge or a combination of both; and this lack of knowledge is particularly important because any dip below quite a high level of knowledge has a devastating effect on the structure and coherence of attitudes to integration4. These features of the orientations of citizens to the European Union confirm the argument outlined earlier in this chapter that well structured, supportive attitudes commensurate with the current stage of integration have not in fact developed. The key question in the current context is: Do these attitudes have any effect on levels of participation in European Parliament elections?

Sources of participation and abstention

The usual explanation given for low turnout in European Parliament elections is that the Parliament has not got enough power. This explanation tends to be cast in terms of the ”second-order election model”. The essence of this approach is that, in compari-

4 For a more detailed discussion see Sinnott, 1998.

son to national elections, European Parliament elections do not affect the distribution of power and that there is, therefore, less at stake in such elections (Reif and Schmitt, 1981, Franklin and van der Eijk 1996). From this starting point the argument is that turnout will be lower and that those who do turn out to vote will be motivated to do so by national considerations and not by European ones. The problem is that this plausible model receives little or no support when tested against the evidence of voter perceptions, attitudes and behaviour.

Evidence on perceptions of power and perceptions of what is at stake in elections indicates that electorates by and large do not make the sophisticated calculations about differences between the power of national parliaments and of the European Parliament or about differences in what is at stake between the two kinds of elections that the second-order-election model requires. Moreover, to the extent that such differential perceptions exist, they have no discernible effect on the propensity to vote Thus, for example, perceptions of the power of the Parliament have little or no effect on turnout (Table 2). There is simply no difference in reported turnout among those with very low, fairly low, fairly high or very high estimates of the power of the Parliament. What does make a difference is if the citizen doesn’t know anything about the power of the Parliament. Likewise, one can show that relative perceptions of the power of the European Parliament in comparison with the power of national parliaments do not have the anticipated effect on turnout. In fact they have a perverse effect – people who see the European Parliament as having more power than their national parliament are slightly less likely to vote for it (Table 3).

Table 2. Type of electoral participation/absention by perception of the power of the European Parliament, 1994 (non-compulsory countries only) (%)

Table 3. Type of electoral participation/absention by perception of difference the power of European and national parliaments, 1994 (non-compulsory countries only) (%)

Turning to the attitudes that do matter, the affective image of the Parliament influences turnout. Moreover, it seems that not having any image of the Parliament is as damaging in terms of the likelihood of voting as having a negative image. The perceived reliability of the Parliament in making sure that the decisions taken by the European Union are in the interests of ordinary people affects turnout. European party and candidate differentials also have an impact (Table 4).

Table 4. Type of electoral participation/abstention by European party differential, 1994 (non-compulsory voting countries only) (%)

The concepts of party and candidate differentials refer to how much it matters to the individual citizens whether particular parties or candidates win seats in the European Parliament elections. The campaign also matters, at least in terms of turnout. But it matters only if the citizen’s exposure to the campaign is active, that is if he or she reads about it or discusses it with family

or friends or colleagues at work. The evidence is that purely passive exposure, that is exposure to the campaign that is limited to being on the receiving end of advertising, or television or radio coverage, or leaflets, or canvassing does nothing to increase the level of turnout. Thus the campaign is influential but by no means decisive.

In the light of the above, perhaps the most surprising contention of previous research is that attitudes to European integration have no significant effect on turnout in European Parliament elections. This counter-intuitive finding may have arisen partly because turnout in European Parliament elections has been treated as a simple dichotomy – voted versus did not vote – and partly because of the limited range of attitudes to integration investigated as possible sources of participation and abstention. In fact, abstention in European Parliament elections is a fourfold phenomenon depending on whether it is voluntary or circumstantial and on whether it is accompanied by abstention in national elections or is specific to European elections (Blondel, Sinnott and Svensson, 1998: 40–43). Consequently, there is not one dependent variable but four, and the same explanation does not cover all four cases.

Obviously, the key dependent variable is voluntary Euro-specific abstention. The various influences on Euro-turnout cited above cast doubt on the prevailing conclusion that attitudes to European integration do not matter. However, such bivariate relationships do not prove the point. What is needed is a multivariate analysis that controls for all the relevant influences. In fact, a logistic regression analysis using a wide range of contextual and attitudinal variables does show that voluntary Euro-specific abstention is significantly affected by attitudes to European integration, by attitudes to the European Parliament and by attitudes to the parties and candidates in the election and that it is not significantly affected by second-order considerations and calculations (Blondel, Sinnott and Svensson, 1998: 222–236). Given the conventional wisdom of the second-order-election-model, perhaps the most important message to come out of this analysis of turnout in the 1994 European Parliament elections is that attitudes to and perceptions and experiences of Europe and of the European Parliament make a

difference to people’s participation. Contrary to what is often argued, European elections are not simply proxy national elections or appendages to domestic politics. This means that appeals cast in terms of domestic political issues are not necessarily conducive either to mobilising voters or to winning votes. This brings us the wider issue of how to respond to the challenge of increasing turnout and improving the quality of participation in European Parliament elections.

Approaches to a solution

There is no instant solution to the problem of turnout in European Parliament elections but there are piecemeal, gradualist steps that can be taken (a) to facilitate participation and (b) to mobilise it. In terms of facilitating participation, three factors need to be considered. First, the timing of the election: Is it really sensible to pick a weekend in mid-June for an election to a Parliament that does not have a secure base in the hearts and mind of the citizens? The data from 1994 show that absence from home, either for the day, for the weekend or for holidays was a significant source of circumstantial abstention (Blondel, Sinnott and Svensson, 1998). At a minimum, the choice of mid-June does not facilitate the voters. Secondly, the evidence suggests that registration and voting card problems lead to a significant amount of circumstantial abstention. Among the non-compulsory voting countries, the problem is most prevalent in Britain, France and Spain but it is also discernible in other countries. Improved administrative arrangements or improved communication concerning existing arrangements would help to solve this problem. Thirdly, there is the issue of the day of voting. If the question is whether voting should be on a weekday or on a Sunday, it is clear that Sunday voting is more conducive to turnout. The evidence suggests, however, that both weekday and Sunday voting inhibit participation, each in different ways. The logical solution would be to allow voting on a Sunday and a Monday. Although this would considerably increase the administrative costs of holding elections, it ought to be seriously considered and not just for European elections.

Turning to the challenge of mobilising the voters, one should note that compulsory voting is a form of voter mobilisation; the problem is that it is coercive mobilisation and, as such, it is inconceivable for European Parliament elections. Holding concurrent elections is also a form of mobilisation; the problem is that it mobilises voters on a sub-European basis and, in itself, does nothing for the quality of European participation. Other institutional factors directly affecting participation include the nature of the electoral system, though the effect of this is more complex than previously thought. More important, however, is the indirect effect that stems from the constitutional system of the EU and from the role of the Parliament within it. Comparing the European Parliament to national parliaments, some have argued that the only way to secure widespread participation in European Parliament elections is to have the European executive elected by and responsible to the European Parliament. European elections would then produce a party or combination of parties with a mandate to govern Europe. Such a proposal must be examined in the light of the nature of power and of the style of decision-making in the European Union.

Power can be concentrated or it can be dispersed. Examples are the traditional British model on the one hand and the American model on the other. But political systems are further differentiated according to whether the style of decision-making is adversarial or consensual. Combining these two dimensions gives four types of democratic governance (see Figure 2).

Figure 2. Types of democratic governance

Concentration/dispersal of power

Concentrated Dispersed


Type I

Type II

Britain The United States

Style of decision-



Type III

Type IV

The Netherlands Switzerland

The point is that democratic governance in the European Union falls into Type IV and is virtually certain to continue do so for the foreseeable future. This must be taken into account in considering measures to mobilise voter participation. Mass mobilisation is not going to be brought about by some grand transformation in the role of the Parliament that would enable European Parliament elections to produce a mandate to govern Europe. Given that a big-bang approach is not on, it is perhaps fortunate that the raw material for a gradualist approach to mobilisation exists. Even in the present state of the EU constitution and of the role of the Parliament, the evidence indicates that the image of the Parliament matters, that the reliability of the Parliament matters, that European party and candidate differentials matter and that overall attitudes to European integration matter. These are all things that can be worked on here and now without waiting for someone or something to transform the European Parliament in ways would not be consistent with the nature of either the institutions or of the political processes of the Union.


Blondel, Jean, Richard Sinnott and Palle Svensson (1998) People

and Parliament in the European Union; Participation, Democracy and Legitimacy, Oxford, Oxford University Press

Converse Philip E. (1964) ‘The nature of belief systems among

mass publics’ in Ideology and Discontent, ed. D.A. Apter. New York, Free Press Eichenberg, Richard, C., (1998) ”Measurement matter: Cumula-

tion in the study of public opinion and European integration”, Paper presented at American Political Science Association Annual Meeting, September 3–6, 1998, Boston Eurobarometer, (1995) Report on Standard Eurobarometer 43,

European Commission Directorate General X, Autumn Franklin, Mark and Caes van der Eijk (eds.) (1996) Choosing

Europe? The European Union and National Politics in the Face of Union, Ann Arbor: University of Michigan Press

Lindberg, L. N. and Scheingold, S. A. (eds.) (1970) Europe's

Would-Be Polity: Patterns of Change in the European Community, Englewood Cliffs, NJ: Prentice-Hall

Niedermayer, Oskar, (1995) ”Trends and Contrasts” in Oskar

Niedermayer and Richard Sinnott (eds.), Public Opinion and

Internationalized Governance, Oxford: Oxford University

Press Reif, Karl-Heinz and Hermann Schmitt (1980) ”Nine second-

order elections: a conceptual framework for the analysis of European elections results”, European Journal of Political Research, 8: 3–4 Sinnott, Richard (1998) ”Knowledge and the position of attitudes

to a European Common Foreign and Security Policy on the real-to-random continuum” Paper presented at American Political Science Association Annual Meeting, September 3–6, 1998, Boston

Integration and Statehood in the European Union

The Distribution of Powers Between the EU and the Member States

Brigid Laffan1


There is considerable scholarly debate about the challenges facing the contemporary nation state and about the changing nature of statehood in Europe, and this debate is focused in particular by two processes – the ending of the Cold War and the evolution of European integration. The continuing centrality of nation states to political order is asserted trenchantly by many authors, who point to the depth and breadth of the institutions they have spawned. Others point to the arrival of more porous borders and to perforated sovereignty, to the possibility of ”prismatic” patters of political authority scattering allegiances and dispersing them not only at the national level but also at the subnational and supranational level, and to multiple loyalties. Controversy about the role of the nation state is bound up with debates about the relationship between European integration and statehood. (Laffan, 0’Donnell, Smith 2000, Moravcisk 1998, Sweet Stone et al. 1998.)

The objective of this paper is to analyse the distribution of powers between the EU and the member states, but in a manner which reflects on the ties and tensions that are moulding the kind of political and economic order that is emerging in the Union.

1 The author is Jean Monet Professor of European politics at Department

of politics at university college Dublin.

Three ”tensions” in integration have a considerable impact on the relationship between statehood and integration and between the member states and the collectivity. These are:

  • the tension between a union of states and a market
  • the tension between the EU as a polity and a problem-solving area
  • the tension between EU level policy-making and policy making at other levels.

These tensions throw up contradictions that have a particular bearing on the subject matter of this conference, namely, the democratic fabric of integration and the participation of citizens in EU politics.

A Union of States and a Market

Promoting and facilitating economic exchange was central to the integration effort from the outset. Patterns of commercial exchange and the perceived economic needs of member states were critical in the development of the common market and the internal market. (Moravcsik 1998, Milward et al. 1993, Laffan et al. 2000). The Union’s economic order operated at the micro rather than the macro level, injecting considerable competitive pressures into the domestic economies and thereby fundamentally altering patterns of production, distribution and exchange. The internal market programme, launched in the mid-1980s, was transformative in character in that it required a strengthening of the political-economic authority of the Union. The removal of internal barriers – physical, fiscal and technical – among national economies reflected a search for greater efficiency and competitiveness in the European economy. It involved a deepening of the common market in goods and an extension of cross-border competition to services and public procurement. Large sections of the national economies that had been relatively untouched by market integration, notably financial services and public utilities, were brought within the remit of the Union’s regulatory range. A whole new language of regulation,

”home country control”, ”mutual recognition”, ”a level playing field” built a new approach to EU regulation. ”The quantitative accumulation of new European laws and market opening measures produced a truly qualitative shift in the dynamic of economic integration and in European economic governance. The transformational nature of the internal market programme led to changes that have a direct bearing on the concerns of this conference. The most significant are:

  • a step change in the regulatory reach and capacity of the EU, which is reflected in the concept of the EU as a ”regulatory state”. This meant more EC laws and the development of new European agencies. EC regulation became more visible and contested within the member states and impacted more directly on individual citizens.
  • the internal market reduced the boundary control capacity of national governments in a variety of ways.
  • the success of the programme generated pressures for further economic integration< notably, the single currency project.
  • the extensive mobilisation of national actors in the Brussels since the mid-1980s, all seeking voice and influence in this growing arena of public policy making. Domestic social forces were drawn out of the national into the European arena of policy making as they beam stakeholders in EU policy making.

The significance of market/economic integration in the process of European integration shapes in a fundamental manner the kind of political and economic order that is emerging in Europe. This is illustrated by analysis of the second tension, highlighted above, notably between the EU as a problem-solving arena and as a polity.

Problem Solving Arena and Polity

Much of the day to day politics and public policy making of the European Union is characterised by the search for solutions to the problems generated by the attempt to make policies that must stretch across such a wide variety of member states. The actors

engaged in EU policy making are purposeful. The Union’s policy regimes have ”been created and sustained by political actors, acting purposively to achieve specific goals or to resolve specific problems”. (Wallace, 2000, p. 7.) They must be convinced of the benefits of collective action and the appropriateness of the Union as the arena in which to tackle these problems. They are engaged in a variety of projects, not necessarily the project of European integration per se. The multileveled system of policy-making has produced multiple arenas and a diverse set of policy subsystems rather than a traditional political hierarchy. EU public policy making is non-hierarchical, characterised by ongoing negotiations and segmented in different institutional settings and policy networks. The technical nature of much of what the EU does privileges the politics of the pragmatist and the experts who populate the myriad of committees attached to the services of the Commission and the Council sub-structure. Much of what is agreed occurs at the level of the specialists rather than their political masters. The complex nature of the Union’s institutional landscape and the multiple decision rules that animate the policy process privileges ”insiders”, those with knowledge and experience of how the EU process works. To the hundreds of thousands of officials and representatives of interest organisations who populate the Brussels level arena, ”Euro” politics is familiar both in terms of substance and process. To the mass electorate, on the other hand, weakly connected to Euro politics, the EU is remote. arcane, and unwieldy. Paradoxically, it is precisely those attributes of the system that endow it with an authoritative role in public policy making raise difficult issues of polity.

The deepening of market integration, the intensification of treaty change and big projects such as EMU and enlargement brought polity issues – rather than just the ”politics of policy” – to the fore. Up to the early 1990s, the authority and legitimacy of the EU could rest on the ”shadow of the past” and the instrumental benefits it brought to the member states and their peoples. This appears to be insufficient as the EU goes beyond market regulation in the narrow sense. European integration has become politicised and is a divisive issue in the domestic politics of many states.

Lines of resistance to integration appear to be on the increase albeit with considerable cross-national variation (Wallace, 2000). A marked characteristic of scholarly writing on the EU in the 1990s is a focus on the so-called ”democratic deficit”. The term deficit implies an inadequacy and insufficiency, but what is the appropriate benchmark against which to test the Union? To date the process of enhancing the democratic element in the Union is following the well-worn Union path of incremental change and pragmatic adaptation, not unlike the process of market integration. However, unlike market integration, there is no over-arching goal or big idea. In fact, national political leaders are extremely cautious in relation to the politics of integration. They are far happier to launch big projects in the economic sphere than to disturb the balance or logic of domestic politics. The same political leaders show a marked reluctance to communicate the realities of power in contemporary Europe to their electorates. They persist with the old language of national interest when Janus-like they serve both the national and collective European governance. The realities of interdependence in Europe and more widely mean that European governments must embrace a broad agenda of co-operation. This leads to the next tension, the tension between national and European policy making.

Tensions in Multileveled Policy Making

All multileveled systems of public policy making are characterised by tension and contention about the appropriate level of policy making and implementation. The Union, as a multileveled polity, is no stranger to such conflicts. I was asked specifically in this section to address the distribution of powers between the national and the European. However, it is not just a question of the amount of EU competence in any one area but the overall pattern of economic governance that results from the manner in which the EU engages in public policy making. The pattern since the 1950s is one whereby more and more areas of public police have come within the ambit of collective governance. The major constitutional settlements, the treaties, have been critical in the formal extension of

EU competence. The treaties reflect a dual process. On the one hand, new treaties formalise policy co-operation that was already taking place. This is particularly true of EU R&D programmes, environmental policy, education and culture, co-operation on justice and home affairs. On the other hand, the treaties may set up new policy goals such as the single currency by defining the calendar, processes, institutions, and eligibility criteria for membership of the new currency zone. EU competence has expanded in other ways, notably, by the judicial activism of the Court of Justice and the ”policy entrepreneurship” of the Commission. But the treaties tell us that the EU has few exclusive competencies, apart from the common commercial policy, elements of competition policy and more recently monetary policy for the ”Euro 11”. EU public policy making operates on the basis not of exclusive competence but more on the sharing of competence with national authorities. This has led to what has been termed ”creeping competence” whereby the EU now has a role, however, minor, in many areas of national policy. The nature of the Union’s role differs depending on the kind of policy.

As noted above, the Union has formidable power in the regulatory field because of the common-internal market process. This has lead to very significant law making to provide a regulatory framework for the free movement of the factors of production. It ensures that law is the main source of public power in the Union. EU regulation has gone beyond the narrow confines of market regulation to include elements of social and environmental regulation. Given the importance of the market to the member states and economic actors within them, we are unlikely to see a significant reduction in the Union’s regulatory capacity. In fact concern about food safety, for example, may lead to more not less EU regulation. The focus now is on improving the regulatory regimes and in ensuring the EC law is transposed and enforced in the member states. There is also an important international dimension to EU regulation arising from demands for global regulation in certain fields.

Redistribution remains a largely national responsibility in Europe. That said, the Union has had to develop a re-distributive

capacity to facilitate the process of market integration and to provide side payments in return for treaty change. Prior to 1988, the common agricultural policy was the main re-distributive policy in the Union. Since then it has been augmented by the structural and cohesion funds that channel finance to the less developed member states and regions in the Union. There was a doubling of the amount of EU finance going to Europe’s poorer regions between 1988–1992. This was followed by further increases in the next period to 1999 and a stabilisation of EU cohesion spending to 2006, following the agreement on a new financial perspective at Berlin (24–25 March). More money was accompanied by new processes for distributing EU finance. An expansion in re-distributive policies was accompanied by a number of distributive policies in research and development, education and culture. The small size of the EU budget places a clear limit on the capacity of the Union to greatly expand its distributive and re-distributive capacity. The Berlin agreement retains 1.27 percent of GDP as a cap on EU finances.

The Treaty on European Union in 1992 expanded the ambitions of the Union in a number of directions, notably the common currency, the common foreign and security policy and co-operation in justice and home affairs. The single currency has been established and was accompanied by institution building in the form of the central bank and the stability and growth pact to govern national budgetary policies. The management of the single currency has greatly increased the need for the member states to develop common understandings of economic developments both within the Euro zone and in the international monetary system. Supranational influences are likely to increase and the range of issues that are of ”common concern” to the member states is likely to grow. The debate has already moved to fiscal policy and issues of unfair tax competition. The activities of the Union in pillars two and three have grown in importance and will become even more central to the agenda of the Union. The war in Kosovo has re-opened the debate about the Union’s international capacity and its need to develop an enhanced capacity in relation to European security. The Treaty of Amsterdam established a very ambitious programme in

the area of justice and home affairs, which may strengthen the modest framework for collective action that now exists.

The Balance between the National and the European

The EU system is a system that works on the basis of a high level of decentralisation and delegation. It is not a system with a strong centre replete with large fiscal or bureaucratic resources. It works on the basis of the enmeshing the national and the European in an uneasy tension between integration and autonomy. With the establishment of the single currency, the EU has put in place the last building brick of economic integration. Enlargement to the east and south will eventually establish the geographical reach of the system and one or two more intergovernmental conferences will settle the main institutional questions. The system may well be entering a period of consolidation following the dynamic of change since the mid-1980s. This should allow the relationship between EU competence and national competence to settle down. The decision at Amsterdam to insert the protocol on subsidiarity and proportionality into a treaty format makes these principles part of the constitution of the system. The protocol establishes how Article 3b in the Treaty on European Union should be operationalised. The document is beautifully ambiguous. It copperfastens subsidiarity as a legal principle but one that ”does not call into question the powers conferred on the European Community by Treaty, as interpreted by the Court of Justice”. The application of the principle must also respect the provisions and objectives of the Treaty ”particularly as regards the maintaining in full of the acquis communautaire and the institutional balance” (Protocol, articles 2– 3). Moreover, it identifies subsidiariry as a principle that ”allows Community action within the limits of its powers to be expanded where circumstances so require, and conversely, to be restricted or discontinued where it is no longer justified” (Protocol 3). Provided that these provisions are met, the subsidiariry principle is one that should govern all EU institutions in the carrying out of their tasks. The protocol was mindful that the EU should legislate only

to the extent necessary and that Community measures should leave as much scope for national decision as possible. Moreover, ”care should be taken to respect well established national arrangements and the organisation and working of Member States” legal systems. (Protocol, article 7.) The debate about subsidiarity and its subsequent inclusion in the treaties reflects the fact that the Union is now maturing as a public policy arena and that care must be taken to ensure that the Union’s reach does not extend beyond its grasp.

The Balance between Large and Small States

The constitutional order of the Union is based on the formal equality of the 15 member states. That said, the member states differ in size, economic wealth and capacity. To date, the EU has managed to balance the needs of its large and small states are a relatively smooth manner. Size does not feature as a variable in the day to day policy-making system of the Union. Policy cleavages would tend not to manifest themselves as cleavages between the large and small. Small states do, however, have a collective interest in issues of representation, voice and the rules of the game. The current system is one that gives over-representation to small states in terms of weighted votes in Council and MEPs in the European Parliament. The prospect of eastern enlargement with many more small states has raised the question of small/large state representation, an issue that was not resolved at Amsterdam. In a protocol on institutions to the Amsterdam Treaty the outline of an institutional bargain is there in that the large states will lose a second Commissioner in return for a re-weighting of votes in the Council. This will be one of the most imponant issues to be addressed at the 2000 Intergovernmental Conference. Apart from the question of formal representation in the institutions, smnll states must be vigilant about the emergence of an informal directoire of the large states in the Union. Large states already have more extensive bilateral contact with each other than they do with the small states and

there has been a growth of informal contact groups in the second pillar. High-ranking posts in the Union have also tended to go to the nationals of large states in the last year.


The Union has achieved deep economic integration without a centralised political authority. It has developed a capacity to govern this increasingly integrated economic space without many of the attributes we associate with governance. Collective EU governance had to be crafted in a novel and experimental manner because of the continuing salience of the member states as containers of national societies and polities. In general, the EU does not represent a process of zero-sum bargaining between the national and the European. The EU can only succeed by mobilising the institutions, actors and resources of the member states and therefore by reaffirming them. In the EU, states trade formal sovereignty and control for representation, voice and influence. This should be seen as a positive sum game, although not for all states and all social forces at any one time. The prismatic nature of the Union is set to continue with segmented bargaining and problem solving in different policy arenas. Arising from EMU, the system will have to take on a more domestic and less diplomatic character. The trajectory of the Union as a polity is very difficult to plot, as the politicisation of the EU system is novel and the emphasis on bringing Europe closer to its citizens is a relatively recent concern for Europe’s political leaders. It could be argued that the Union could be democratised in bits and pieces in line with the way in which the Union has done things in the past. At present, the Union’s political fabric is being fashioned in an incremental, contested and patchy manner. The process is not unlike a strategy of pragmatic and incremental market creation. However, the disquiet in a number of member states about the impact of the EU on their national democracies, in addition to the growing stakes in integration may force Europe’s political leaders to be more innovative in their

approach to the political fabric of integration. So far they have resisted tampering with the ”roles of the game” of national politics.


Laffan, B., O’Donnell R. and Smith, M. (2000, Europe’s Experi-

mental Union: Rethinking integration. London: Routledge

Milward, A. S., Lynch, F.M.B., Romero, F., Ranieri, R. and Soren-

sen V. (1993) The Frontier of National Sovereignty. London: Routledge Moravesik, A. (1998) The Choice for Europe. Ithaca: Cornell Uni-

versity Press Sandholtz, W. and Stone Sweet A. (1998) European lntegration and

Supranational Governance. Oxford: Oxford University Press

Wallace, H. (2000) ”The Policy Pendulum”, in Wallace, W. and

Wallace, H., eds. PolicyMaking in the European Union. Oxford: Oxford University Press

The Architecture of EU Institutions and Citizen Participation

Hermann Schmitt1


Modern democracy is representative democracy. This holds for the prototype of mass democracy, that is the political process in the national political arena. It holds all the more for a political entity which is as large and complex as the European Union. Some 300 million citizens are called to participate in all sorts of local and regional politics, in national politics and in EU politics. The different tiers of this multi-tiered polity are not independent from one another, and the national and the EU layer are particularly intimately intertwined.

Direct citizen participation in government is generally limited. At most places and most of the times, political decisions are taken by representatives of the people rather than by the people themselves. Citizen participation in politics is not restricted to, but concentrated upon the process of electing representatives. Political parties are in many ways indispensable in this process.

Even if we leave aside possible irregularities, evaluations of the democratic quality of the electoral process may vary from one election to the next. Those evaluations can be based upon two

1 The author is professor in political science at Mannheimer Zentrum für

Europäische Sozialforschung at Mannheim university.

criteria. One is turnout. Do the people who are not in a position to take part in the real thing (i.e., policy making) at least participate in elections of their representatives, and what does it mean if many do not? The second criterion is political agreement (or, more technically, issue congruence) between electors and elected. Representative democracy works well if voters participate in elections, and if representatives decide as voters would if they had a chance to. A related question is how issue congruence is brought about. Are the parties responsive followers of voters preferences or are they powerful molders of the world views of their voters. We will briefly address each of these questions in turn.


Turnout is high in national parliament elections, with only a very modest downward trend over the last half century. It is considerably lower in European Parliament elections, with a dramatic decline in the last election (Figure 1).

Figure 1. Electoral Participation in Western Europe and in the EC/EU

There are two aspects of turnout differences between national and European Parliament elections: level and trend. Level differences are mainly due to differential mobilisation and involvement. It is more at stake in national elections than in European elections, the campaigning of political parties is more intense, the politicisation of the mass public is generally higher, and the public is more involved as a result (Reif & Schmitt 1980).

It is generally believed that level differences do not relate systematically to attitudes about the EU. EU-critical attitudes are not a major factor in the explanation of abstentionism in European Parliament elections (see Schmitt & Mannheimer 1990 and van der Eijk & Franklin 1996; for a contrary view Blondel et al. 1998). People do not abstain because they do not agree with the Union and its policies, but because they are uninterested and disenchanted. Rather than abstentionism, one would expect EU-critical orientations to cause anti-integrationist parties or lists to form and to absorb the respective vote share (as, e.g., in the Danish and the French case). The formation of such EU-critical partisan formations may take time, and it may not be equally successful in every party system. Nevertheless, the general impression is that abstentions in European Parliament elections are not normally a result of EU-critical orientations.

It is not just level differences which become apparent in the above figure. There are also trend differences: participation rates in European Parliament elections decline faster than participation rates in national first order elections. Why this is so is not well understood; changing context effect patterns like concurrent national first-order elections in different national settings may be one reason; composition effects that originate in the two rounds of enlargement of the Union (since 1979) may be an other. This in any case needs further investigation.

While low turnout is best characterised as a ringing alarm clock for the democratic quality of elections, deficient representation is immediately and substantially more important. Low turnout is not per se problematic – those staying home may be well in agreement with the functioning of the political system and may therefore see no need to participate. In the long run, however, deficient

participation seems likely to undermine effective representation. This is the question we now address: Whether and to what degree this already is apparent in the European Union.

Effectiveness of Political Representation

The European Union combines two distinct modes of government which are based on two different routes of socio-political linkage: intergovernmentalism and the confederal model of political representation and party government and the federal model of political representation (see Figures 2 and 3).

The democratic quality of the EU political process cannot be adequately evaluated if we restrict our attention to just one of the two competing models of government in the European Union, the federal or the confederal. The same goes for the two electoral arenas involved, the European and the national. European Parliament elections are just one, and not the most important, electoral event upon which an analysis of the democratic quality of the EU political process has to be based.

Figure 2. The intergovernmental model of European political representation

European Council/

Council of Ministers

National Government 1 … National Government 15

National Parliament 1 … National Parliament 15

National Parties … National Parties

National Electorate 1 … National Electorate 15

Figure 3. The federal model of European political representation

European Government

European Parliament

European Political Parties

European Electorate

Previous research has established that the effectiveness of political representation in the EU does not differ much either way (Schmitt & Thomassen 1999, chap. 9). Representation functions equally well or badly via the European (federal) as it does via the national (inter-governmental) channel of EU government. It is satisfactory with regard to the grand directions of public policy. While party elites are generally somewhat more to the left than their voters, the congruence between elites’ and voters’ ideological orientations is still considerable. (Figure 4).

Figure 4.

Effectiveness of representation regarding grand policy

directions: left-right orientations

Source: European Representation Study 1994–7.

The political agreement between voters and party elites is equally profound regarding the principle of European unification though party elites are generally somewhat more Euro-positive than their

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9


vo te rs ’ m ea n po sit ions


left right


MEPs’ mean positions


voters are. Elites and voters of most parties meet in the Europositive angle. Only a few find one another in the Euro-critical corner. Mass-elite agreement is found less satisfactory only when we leave the grand political avenues and take the thorny paths of specific EU policy measures. No matter what EU policy domain we look at – open borders, single market, or common currency – the effectiveness of political representation is here found to be deficient. Figure 6 displays the common currency issue as an example. Most party elites are strongly in favour of it, while their voters locate themselves more ”middle-of-the-road”. This may be understood substantively as a more sceptic and critical view. It can however be that many voters feel poorly informed and are therefore undecided about rather than opposing the European Monetary Union.

Figure 5. Effectiveness of political representation regarding specific EU policies: The currency question

Source: European Representation Study 1994– 7.

Political representation is less effective in view of more specific EU policies (and probably in view of specific policies in general) while it seems to work well regarding the grand policy directions – like

left or right, and pro- or anti-integration. Having said this, however, we still know nothing about how this solid base of political agreement between representatives and represented is brought about? Is it the elites that lead voters opinion, or are the voters the driving force?

This can only be adequately assessed with a diachronic study design. Analysing voters’ and party elites’ average political preferences with regard to European unfication in 1979 and 1994, we find some initial evidence that it is the party elites who behave responsively in view of changing voter preferences (Figure 6). Somewhat in contrast to the findings of Essaiasons and Holmbergs Swedish study (1996) this seems to suggest that political representation in Europe might not be such an elite-driven process after all.

Figure 6. Representation from above? A simple causal model linking elites and voters views about European unification in 1994 and 1979

Source: European Election Studies 1979 and 1994.

Notes: N of cases (parties) = 18. The parties for which elite and mass data are available for both 1979 and 1994 are: CVP (B), SP (B), PVV (B), SD (DK), V (DK), FbmEF (DK), CDU (D), SPD (D), PS (F), UDF (F), MSI/AN (I), DC/PPI (I), CDA (NL), PVDA (NL), VVD (NL), Conservative Party (GB), Labour Party (GB), and Liberal Party/Liberal Democrats (GB). The model has been estimated with EQS; goodness of fit measures and test statistics are not reported as there was no intention to maximise the fit between model and data.

Summary and Conclusions

1. Compared to national politics, participation of EU citizens in EU politics is poor and declining. This should not be taken as an indication of political disagreement between party elites and party voters on EU policies because abstention can have, and are repeatedly found to have, other sources than policy disagreement. Factors conducive to high turnout levels are:

high politicisation of the campaign, i.e. high saliency of issues. This is hard to achieve in a European Parliament election as EU issues are typical low salience issues in most places, and there is hardly any disagreement between the major parties about them. successful personalisation of issue/goal conflict. This again is hard to achieve as, for the time being, there is no transnational personalisation of political alternatives at the EU level. One procedural measure that could change this so some degree would be the direct election of the President of the European Commission in concurrence with the EP election. the closeness of the electoral race, i. e. the likelihood of governmental change. This does presently not apply to European Parliament elections, as they do not now result in government formation. However, the investiture of a new Commission by a newly elected European Parliament could be one step in this direction. numerous and strong party attachments/party loyalties. Quite generally, party attachments tend to decline. In this perspective, declining turnout rates in European and national elections are a result of the same socio-political processes. technical measures like compulsory voting, Sunday voting and ”easiness” of the electoral system more generally, and the distance to 1st order election day. These factors tend to explain variance in turnout between countries.

2. Political representation is found to be well-functioning in view of the grand directions of public policy, both generally and in view of European unification. It is found to be ineffective in view of EU policies both in European and in national elections. Thus we conclude that:

national elections are as poor a mechanism for generating voter-party agreement on EU policies as European Parliament elections are (and European Parliament elections are as good a mechanism for common concern issues as national elections are). elections generally seem to translate voter preferences about the grand directions of public policy, rather than those about specific policies, into elite attitudes and behaviours.


Blondel, Jean, Richard Sinnott and Palle Svensson (1998) People

and Parliament in the European Union. Participation, Democracy, and Legitimacy. Oxford: Clarendon Press

Esaiasson, Peter and Sören Holmberg (1996) Representation from

Above: Members of Parliament and Representative Democracy in Sweden. Aldershot: Dartmouth

Reif, Karlheinz and Hermann Schmitt (1980) ”Nine national

second-order elections: A systematic framework for analysis of European elections results”; together with K. Reif), European

Journal of Political Research 8,1: 3–44

Schmitt, Hermann and Jacques Thomassen, eds. (1999) Political

Representation and Legitimacy in the European Union. Oxford:

Oxford University Press 1999 Schmitt, Hermann and Renato Mannheimer (1991) ”About Voting

and Non-Voting in the European Elections of June 1989” in

European Journal of Political Research 19,1:31–54.

Van der Eijk, Cees, Franklin, Mark N., et al. (1996) Choosing

Europe? The European Electorate and National Politics in the Face of Union. Ann Arbor: Michigan University Press


Blaming the Messenger?

Political Communications and Turnout in EU Elections

Pippa Norris1

The erosion of turnout evident in the June 1999 elections to the European Parliament has set widespread alarm bells ringing in Brussels, Strasbourg and Luxembourg as further evidence that the public is becoming disenchanted and disengaged with the European Union. The level of voting participation fell from almost two-thirds (63 %) of the electorate in the first direct elections in 1979 to just under half (49.2 %) of European citizens in June 1999, its historical nadir (see Figure 1). The decline over successive election is particularly clear in the Netherlands and Portugal, as well as in Austria, Finland and Sweden, which saw a sharp drop after their first ”founding” European elections. There are also stark national differences, in the most recent elections 90 % of Belgian citizens voted compared with only one quarter (23 %) of the British electorate (see Table 1). As Franklin and his colleagues have argued2, the institutional system of electoral laws provides much of the explanation for these persistent systemic differences, notably the use of compulsory voting (in Belgium, Luxembourg

1 The author is professor at John. F. Kennedy School of Government at Harvard University.2 Mark Franklin, Cees van der Eijk and Erik Oppenhuis. 1996. ”The Institutional Context: Turnout”. In Choosing Europe? The European Electorate and

National Politics in the Face of Union, edited by Cees van der Eijk and Mark

Franklin. Ann Arbor, MI: University of Michigan Press.

and Italy), Sunday polling day (Austria, Belgium, Finland, France, Germany, Greece, Italy, Luxembourg, Portugal, Spain and Sweden), and the majoritarian first-past-the-post electoral system (used in mainland Britain until the 1999 contests).

Figure 1. Turnout EU Elections 1979–99

Set in a broader context, the decline in the EU turnout is even more worrying because it proves contrary to trends elsewhere. Many assume that there has been a general decline in voter participation but in fact in established democracies levels of turnout have remained fairly stable during the last two decades; 71 % of voting age population participated in elections in these states in

the 1990s, down only 3 % from the 1970s3. As is well known, despite similar socioeconomic and political developments in postindustrial societies, there are persistent cross-national disparities in levels of electoral participation. Some countries like Switzerland, France and the Netherlands have experienced substantial longterm falls (see Figure 3). In American presidential elections, in 1996 less than half (47.2 %) the voting age popular cast a ballot, down from almost two-thirds (63.1 %) in 1960. If we calculate average turnout in 171 countries worldwide, in all national elections from 1945 to 1998, Switzerland ranks 137th, the US ranks 138th, and Mexico ranks 139th.

Figure 2. Change in EU Turnout 1994–99

But, within this context, can the news media or party campaigns be blamed for the downward trend in European participation? If there were a systematic negative bias in the media – if news headlines

3 IDEA. Voter Turnout from 1945 to 1998.

highlighted corruption at the Commission, bureaucratic over regulation in Brussels, and junkets for the European parliament – it would not be surprising if voters deserted the polling stations.

Table 1. Turnout in European Elections, 1979–1999

To examine this issue the first section of this paper outlines the theoretical framework in the literature, the next considers the evidence for the association between attention to campaign communications and turnout in European elections in 1989, 1994 and 1999. The conclusion briefly summarizes the theory of a virtuous circle to explain the major findings and speculates about the implications for participation in EU affairs.

Figure 3. Postwear Mean Turnout Per Decade

Understanding Political Participation

Explanations of political participation focus on four sets of factors. The institutional perspective stresses the importance of the legal context including the level of political rights and civil liberties, the type of electoral system, the facilities for registration and voting, the expansion of the franchise, the frequency, level and timing of elections, and the competitiveness of electoral politics4.

4 See Ivor Crewe. ”Electoral Participation.” In Democracy at the Polls, edited by Austin Ranney and David Butler. Washington, DC: AEI Press; Arend

In one of the most thorough comparative studies, Jackman and Miller examined voter participation in twenty-two democracies and found that political institutions and electoral laws provided the most plausible explanation for variations in voter turnout, including levels of electoral participation and proportionality, multi-partyism and compulsory voting5. Franklin, van der Eijk and Oppenhuis argue that variations in participation in European elections can be attributed in large part to differences in systemic factors, notably the use of compulsory voting, the proportionality of the electoral system, and the closeness of European to nation elections6. In the United States, as well, the legal hurdle of registration requirements and the frequency of elections are widely believed to depress American turnout7.

At individual-level, the cultural perspective based on survey analysis has emphasized the importance of individual resources, like education, age, socioeconomic status and time, combined with


meaning the attitudes people bring to the electoral

process like a sense of efficacy, political interest and party identification. Almond and Verba stressed the importance of ”civic values” learnt through the early socialization process8. Cultural attitudes towards the political system vary substantially across European states, notably support for the regime and representative

Lijphart. 1997. ”Unequal Participation: Democracy’s Unresolved Dilemma.”

American Political Science Review. 91: 1–14.

5 Robert W. Jackman and Ross A. Miller. 1995. ”Voter Turnout in the Industrial Democracies During the 1980s.” Comparative Political Studies, 27: 467– 92. See also Richard Katz. 1997. Democracy and Elections. Oxford: Oxford University Press.6 Mark Franklin, Cess van der Eijk and Erik Oppenhuis. 1996. ”The Institutional Context: Turnout”. In Choosing Europe? The European Electorate and

National Politics in the Face of Union, edited by Cees van der Eijk and Mark

Franklin. Ann Arbor, MI: University of Michigan Press.7 Raymond Wolfinger and Steven Rosenstone. 1980. Who Votes? New Haven: Yale University Press.8 Gabriel Almond and Sidney Verba. 1963. The Civic Culture: Political

Attitudes and Democracy in Five Nations. Princeton, NJ: Princeton University


institutions9. In a long series of studies, Verba has demonstrated how various forms of participation make different demands of skills, money or time, so that political participation can best be understood as a multidimensional phenomenon10. That is, people who regularly donate money to campaigns, or contact their representative, are not necessarily involved in other dimensions like party work, or community activism. There are different costs and benefits associated with different types of participation. The main categories distinguished by Verba and his colleagues concern voting, campaign work, communal activity, and contact specialists. In addition a few citizens are active across all dimensions, while some are involved in none.

Lastly, the organizational perspective has stressed the role of

mobilizing agencies

, referring to the electoral functions of party

and candidate organizations, group networks like churches, voluntary associations and trade unions, social networks of families, friends and colleagues, and the role of the news media11. Putnam has argued that the decline of dense networks of local associations and community organizations has reduced social capital and contributed towards a long-term erosion of American turnout among the post-war generation. Verba found that churches and voluntary organizations provide networks of recruitment, so that those drawn into the political process through these associations develop the organizational and communication skills that facilitate further activity12. In the United States, Aldrich and Wattenberg

9 Pippa Norris. Ed. Critical Citizens: Global Support for Democratic Gover-

nance. Oxford: Oxford University Press.

10 See Sidney Verba, and Norman Nie. 1972. Participation in America: Political

Democracy and Social Equality. New York: Harper and Row; Sidney Verba,

Kay Lehman Schlozman, and Henry Brady. 1995. Voice and Equality: Civic

Voluntarism in American Politics. Cambridge, MA: Harvard University Press.

11 See, for example, Steven J. Rosenstone and John Mark Hansen. 1995.

Mobilization, Participation and Democracy in America. New York: Macmillan.

12 Sidney Verba, Kay Lehman Schlozman, and Henry Brady. 1995. Voice and

Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard

University Press.

suggest that the decline of party organizations, and their replacement by entrepreneurial candidates, has been critical to this process13.

The role of political communications via parties and the news media fall into this latter category of factors. Parties act as mobilizing agencies through direct communications with voters, including traditional grassroots activities such as canvassing, leafleting and contacting voters, as well as holding party meetings and campaign rallies, and using national advertising or party political broadcasts. The news media serve this function through providing information about parties, candidates and policies that can help to crystallize voting choices, and the partisan press, in particular, has long been thought to help reinforce party support. In these activities through positive messages both parties and the news media can serve to increase party and candidate support among electors, and the propensity to turnout, or they can convey negative messages that function to depress participation.

The literature is divided about the effects of media activity. In the traditional ‘Columbia’ model, partisan-leaning newspapers and party campaigns were seen as playing a vital role in reinforcing support and getting out the vote: ”The more that people read about

or listened to the campaign on the mass media, the more interested they became in the election and the more strongly they come to feel about their candidate...Media exposure gets out the vote at the same time that it solidifies preferences. It crystallizes and reinforces more than it converts.

14 The ”Michigan” model conceptualized attention

to political communications somewhat differently, as itself a minor form of activism, instead of an independent factor capable of influencing turnout. This perspective became so influential that it developed into the mainstream view in studies of political partici-

13 John Aldrich. 1995. Why Parties? Chicago: University of Chicago Press; Martin P. Wattenberg. 1996. The Decline of American Political Parties: 1952–

1994. Cambridge, MA: Harvard University Press.

14 Bernard R. Berelson, Paul F. Lazarsfeld and William N. McPhree. Voting: A

Study of Opinion Formation in a Presidential Campaign. Chicago: University

of Chicago Press. Pp. 246–248.

pation, which rarely treated the media as an important causal factor in their models15. A range of more recent studies has credited the media with boosting public participation16.

In contrast, in recent years many popular commentators commonly suggest that the public has become disengaged through negative messages. There are two separate issues here. One concerns the effects of the use of negative or ”attack” ads by politicians where candidate or party campaigns criticize their opponents’ character or record17. In the United States, Ansolabehere and Iyengar provide some of the most convincing experimental evidence that the use of ”negative” or ”attack” television campaign ads, meaning those designed to criticize the opponent, has the capacity to turn off American voters at the ballot box. ”Negative

advertising drives people away from the polls in large numbers… Negative advertising breeds distrust of the electoral process and pessimism about the value of an individual’s own voice.”

18. Yet it is

15 See Angus Campbell, Philip Converse, Warren E. Miller and Donald E. Stokes. The American Voter. New York: John Wiley. P. 92. See also Sidney Verba, Norman H. Nie and Jae-On Kim. 1978. Participation and Political


. Cambridge: Cambridge University Press; Geraint Parry, George Moyser and Neil Day. 1992. Political Participation and Democracy in Britain. Cambridge: Cambridge University Press.16 Brice E. Pinkleton, and Erica Weintraub Austin. 1998. ”Media and Participation: Breaking the Spiral of Disaffection.” In Engaging the Public: How

Government and the Media can Invigorate American Democracy. Ed. Thomas

J. Johnson, Carol E. Hays and Scott P. Hays. New York: Rowan & Littlefield; Hugh Culbertson and Guido H. Stempel III. 1986. ”How media use and reliance affect knowledge level.” Communication Research. 13:579–602; Alexis S. Tan. 1980. ”Mass media use, issue knowledge and political involvement”.

Public Opinion Quarterly. 44: 241–48.

17 See, for example, Kathleen H. Jamieson. 1992. Dirty Politics. Oxford: Oxford University Press; Kathleen H. Jamieson. 1984. Packaging the

Presidency: A History and Criticism of Presidential Advertising. New York:

Oxford University Press; Karen S. Johnson-Cartee and Gary A. Copeland. 1991. Negative Political Advertising: Coming of Age. Hillsdale, NJ: Erlbaum.18 Stephen Ansolabehere and Shanto Iyengar. 1995. Going Negative: How

Political Advertisments Shrink and Polarize the Electorate. New York: Free

Press. P. 112.

difficult to know how far we can generalize from these findings more broadly in part because of the different institutional context of advertising in European campaigns19. Commercial political advertising has come late to most European countries. In some, like Austria, negative advertisements are banned by law. In others like the Netherlands, although ads are allowed, few are aired because parties have limited financial resources. And in still others like Britain it is difficult to compare the effects of a five to ten minute party political broadcast, shown once per channel, with the effects of repetitive 30-second ads common in the United States. Lacking systematic comparative data on exposure to negative ads, we cannot pursue these claims further here.

The other concern claim relates to common practices originating in the news media, which we can examine, such as where routine headlines emphasize political scandals, government incompetence and/or partisan conflict20. For example, Patterson suggests that American voters are turned off by the media’s routine emphasis on the ”game” schema, characterized by horse race journalism (who’s ahead, who’s behind) and extensive coverage of opinion polls21. He argues that changes in journalism in the 1960s produce a shift towards game-immersed news, strengthening voters’ mistrust of the candidates and reducing their sense of involvement. For Cappella and Jamieson strategic frames for political news activate cynical responses to politicians, governance and campaigns22. Yet others argue that a strategic focus and horse race polls function in a positive way, by increasing the American public’s attention to issue information and political knowledge. Zhao and Beske conclude that coverage of opinion polls is complimentary to issue

19 Lynda Lee Kaid and Christina Holtz-Bacha. 1995. Political Advertising in

Western Democracies. Thousand Oaks, CA: Sage.

20 Joseph N. Cappella and Kathleen Hall Jamieson. 1997. Spiral of Cynicism:

The Press and the Public Good. Oxford: Oxford University Press.

21 Thomas E. Patterson. 1993. Out of Order. New York: Vintage.22 Joseph N. Cappella and Kathleen Hall Jamieson. 1997. Spiral of Cynicism:

The Press and the Public Good. Oxford: Oxford University Press. P. 139.

coverage, stimulating rather than displacing attention23. It seems equally plausible that what matters for electoral participation is

what the polls report, not the extent of their coverage per se. In

Britain, for example, Heath and Taylor found that the closeness of the race, as monitored by reported opinion polls, is one of the best predictors of turnout24. Neck-and-neck contests increased the incentive to vote. In addition, the effects of negative news are not well established. In the British context, for example, large-scale experiments in the 1997 election demonstrated that exposure to ”negative” television news about the major parties had no influence on party images or propensity to vote, whereas positive news did have a significant impact on voters25.

Therefore we need to go further to understand the effects of political communications on public participation. Since most of the research has been conducted within the context of American campaigns, which are atypical of most established democracies, it is useful to reexamine the evidence in a broader range of post-industrial societies. This study focuses on political participation in European elections, which allow us to explore the effects of campaign communications across the fifteen member states, controlling for some of the major cultural and the structural factors already discussed. Political participation involves many different types of activity, from contacting representatives to becoming active in community organizations, political parties, or interest

23 Xinshu Zhao and Glen L. Beske. 1998. ”Horse-Race Polls and Audience Issue Learning.” The Harvard International Journal of Press/Politics. 3(4): 13– 34; Philip Meyer and Deborah Potter. 1998. ”Pre-election Polls and Issue Knowledge in the 1996 U.S.Presidential Election.” The Harvard International

Journal of Press/Politics. 3(4):35–43.

23 Denis McQuail. 1992. Media Performance: Mass Communication and the

Public Interest. London: Sage. P. 17.

24 Anthony Heath and Bridget Taylor. 1999. ”Turnout and Registration.” In

Critical Elections: Voters and Parties in Long-term Perspective, Edited by

Geoffrey Evans and Pippa Norris. London: Sage.25 See Pippa Norris et al. 1999. On Message: Communicating the Campaign. London: Sage. Chapter 9.

groups. In this study we focus on comparing voter turnout, one of the least demanding forms of activity but also one of the most universal. For many people, casting the ballot provides their only form of political expression. This measure is also comparable across established democracies, unlike involvement in parties or interest groups that may mean very different things in different institutional settings.

Political Communications and Electoral Turnout

As noted earlier, the level of voting participation fell from almost two-thirds of the electorate in the first direct elections in 1979 to just under half of European citizens in June 1999. Can the news media or party campaigns be blamed for the downward trend in European participation? Theories of videomalaise suggest that those who were most exposed to the news media and/or party messages should prove the most cynical and disenchanted with Europe. Yet if we compare the sources of campaign information for those who reported voting in the 1994 European elections, contrary to the videomalaise thesis, a positive relationship is evident: those who saw something about the campaign in newspapers and television, or who received an election leaflet or saw party advertising, were more likely to cast a ballot (see Table 2). All the zero-order correlations (with the exception of being contacted by a party worker) proved significant. In many cases the gap between voters and non-voters proved modest but in the case of newspaper readers the gap reached 10 percentage points.

Table 2. Political Communications and Voting Participation, 1994

To see whether this relationship held up to multivariate analysis, regression models was run in 1989, 1994 and 1999 predicting voting turnout using the standard structural, attitudinal and nationallevel controls. This includes education, age and income (the latter as a proxy for SES), which have most commonly been found to be associated with turnout, along with political interest measured by propensity to discuss politics. The results in Table 3 confirm that, as many previous studies have found, age and income proved strong predictors of turnout, along with political interest. The younger generation is particularly prone to stay home. In contrast the gender gap in turnout has shrunk over the years to become insignificant while education (measured on a restricted scale) proved inconsistent. National factors also proved important, with below-average reported turnout in Britain, the Netherlands, Portugal and Ireland, a pattern already shown in official aggregate figures in Table 1. As noted earlier, the legal and institutional context, such as the use of compulsory voting, provides by far the most convincing explanation for these national contrasts. After social controls were included, all the forms of political communication proved significant and positive, including use of newspapers, television/radio and party campaign activity. The strength of these factors did vary across these models, in part because of functionally equivalent but different measures of media attention, but all pointed in the same direction. The replication of these models in successive elections increases our confidence in the reliability of the results.

Table 3. Predictors of Voting Participation, 1989–99

We find no evidence for the claim that those most exposed to news coverage during the campaign were demobilized by the experience. Yet interpretations of these findings remain open to interpretation. Does media attention to the campaign (which is sequentially prior) lead to turnout? Or does a general propensity to turnout lead to media attention (because I want to cast my vote, I seek out information about the parties and candidates)? Elsewhere, in a more extended treatment, I argue that the most plausible interpretation of this evidence is that there is a virtuous circle

where watching the news activates existing predispositions and prior tendencies lead people to turn on the news26.

Table 4. Evaluations of TV Campaign News by Turnout, 1989

These issues cannot be resolved here with cross-sectional data but to explore some of the reasons behind this pattern we can look at how people evaluated a series of statements concerning television coverage of the 1989 campaign. Table 4 shows that compared with those who did not turnout, voters were significantly more likely to report that TV coverage showed them where their party stood on Europe, helped make up their mind how to vote, and highlighted party differences. This supports the idea, which Paul Lazarsfeld argued fifty years ago, that the attentive use the information on the news to help crystallize their voting choices. In contrast, nonvoters were more prone to feel that coverage left them feeling confused or bored.

26 See Pippa Norris. Fall 2000. A Virtuous Circle: Political Communications in

Post-Industrial Societies. Cambridge: Cambridge University Press.

Conclusions and Implications

Why should we consistently find a positive link between electoral turnout and attention to the news media? There are three possible answers.

One interpretation is that those who are most predisposed to participate (for whatever reason) could well be more interested in keeping up with current affairs in the news, so the direction of causation could be one-way, from attitudes to use of the news media. This view is consistent with the ”uses and gratification” literature, which suggests that uses of the mass media reflect prior predispositions in the audience: people who love football turn to the sports results, people who invest check the business pages, and people interested in politics read op eds about government and public policy27.

Another answer could be that the process of watching or reading about public affairs (for whatever reason) increases our interest in, and knowledge about, government and politics, thereby facilitating political participation. The more we watch or read, in this interpretation, the more we learn. News habits can be caused by many factors such as leisure patterns and broadcasting schedules: people may catch the news because it comes on after a popular sit-com, or because radio stations air headline news between music clips, or because the household subscribes to home delivery of a newspaper. In this view, the direction of causality would again be one-way, but in this case running from prior news habits to our subsequent political attitudes and knowledge.

Both these views could logically make sense of the associations we establish. One or the other could be true. It is not possible to resolve the direction of causality from cross-sectional surveys taken at one point in time. But it seems more plausible and convincing to assume a two way-interactive process. This conclusion argues that in the long-term, like the socialization process in the family or workplace, there may well be a ”virtuous circle” where

27 Jay G. Blumler and Elihu Katz. Eds. 1974. The Uses of Mass Communications: Current Perspectives on Gratifications Research. Beverly Hills, CA: Sage.

the news media and party campaigns serve to activate the active. Those most interested and knowledgeable pay most attention to political news. Learning more about public affairs reduces the barriers to further civic engagement. In this interpretation, the ratchet of reinforcement thereby moves in a direction that is healthy for democratic participation.

In contrast, the news media has far less power to reinforce the disengagement of the disengaged, because, given the easy availability of the multiple alternatives now available, and minimal political interest, when presented with news about politics and current affairs this group is habitually more likely to turn over, turn off, or surf to another web page. If the disengaged do catch the news, they are likely to pay little attention. And if they do pay attention, they are more likely to mistrust media sources of information. Repeatedly tuning out political messages inoculates against their potential impact. This theory provides a plausible and coherent interpretation of why the press and broadcasters may have a negative bias in coverage of the EU and yet those who pay attention to the news media remain more engaged than those who do not watch or read. What matters here are the characteristics of how people sift, sort and use messages in the news media, rather than simply passively and uncritically absorbing the messages in an over-simple stimulus-response model.

Claims of videomalaise are methodologically flawed so that they are at best unproven, to use the Scottish verdict, or at worse false. As a result too often we are ”blaming the messenger” for more deep-rooted ills of the body politic. This matters, not just because we need to understand the real causes of civic disengagement to advance our knowledge, but also because the correct diagnosis has serious implications for public policy choices. ”Blaming the messenger” can prove a deeply conservative strategy, blocking effective institutional reforms.

If we can generalize more broadly from these results, they suggest that campaign communications may thereby reinforce the division between those who tune in and tune out from public affairs. Some people will have more civic skills, social networks, and interest to find out about events in Brussels or Luxembourg,

and to cast their vote accordingly. The institutional context also does seem important for electoral turnout, and devices like compulsory voting and Sunday polling days can raise levels of participation. But the evidence here suggests that coverage of public affairs in the traditional news media should not be blamed for broader inequalities in educational skills or socio-economic resources common throughout post-industrial societies. The erosion of turnout in EU elections can therefore best be explained by changes in a range of other factors, such as the institutional context for voting, the role of mobilizing agencies such as political parties in activating support, the performance of the EU on major issues like unemployment and economic growth, and the rational incentives to cast a ballot if voters cannot perceive major differences between political parties on many of the core issues facing Europe. There are problems in how the news media covers Europe, in particular a lack of stories about most routine policy matters like regional aid or jobs programmes, combined with a systematic negative bias in the newspaper and television coverage that is provided on issues such as the euro28. Nevertheless often the news media is blamed as the messenger, rather than confronting more deep-rooted problems in how citizens connect with the European Union.

28 See Pippa Norris. Fall 2000. A Virtuous Circle: Political Communications in

Post-Industrial Societies. Cambridge: Cambridge University Press.

To Share Democratic Legitimacy Between Different Political Levels

Hans Agné1

Even though democracy is about power, equally shared between citizens, democratic procedures have never controlled more than parts of the power exercised in any society. To understand democracy we thus study the relation between the unit in which democracy is thought to take place and other units where power is exercised, democratically or not. The question I would like to pose concerns the division of power between the European union and its member-states. Given a preference for high turnout in general elections and its possible effect in democratic legitimacy: Which power-division is preferable? I start by making two assumptions: the first answers the question why there is a general interest in the level of turnout; the second answers the question why citizens vote. I then proceed by discussing how turnout and democratic legitimacy are affected by power-division between more or less democratic organisations, focusing on the European union and its member-states.


1 The author is doctor student at Department of political science at

Stockholm university.

Why are we interested in citizens participation in general elections? A social scientist answer could be that turnout in general elections is a good indicator of the popular support for the political system (or some other politically relevant quality). That answer is fair enough for explaining the interest of a social scientist. However, if the turnout in general elections is interesting only as an indicator of the popular support for the political system, and not in itself important for the popular support, than no one outside social science need to bother about it; the political debate would focus on the real problem, that is a supposed lack of popular support for the political system, and not waste any time on the mere instrument for studying the problem, which is the turnout. Thus, our first answer can not explain the concern for turnout in general elections found in the political sphere, and so we have to put the question again. Why are we interested in citizens participation in general elections?

A second answer would be that democracy is impossible if citizens don’t vote. Even though this answer probably could explain some of the common interest in citizens participation in general elections, it’s not sufficient. If you adopt a representative or elitist conception of democracy – which I believe most people do – voting citizens is not a necessary condition for democracy. What is a necessary condition for representative democracy is the possibility for all citizens to vote in free and fair elections. Of course one could suspect the voting-possibility to be only formal, and not real, if there is an extremely low turnout, as in the European elections, that is to say: citizens are prevented from voting by some invisible structure of power determining society. But such a marxist interpretation would require extra judgement, it can’t be accepted a priori, a judgement which hasn’t been made in the Swedish debate. So the idea that democracy is impossible if citizens don’t vote, can’t fully answer our question why there is a general interest in turnout in general elections, as long as some of us are neither marxists nor participatory democrats.

To understand the common interest in turnout in general elections I would propose a third answer, complementing the former ones, saying that participation in general elections has a common

interest because it’s assumed to improve or underpin the legitimacy of the political system. In a weberian understanding of legitimacy (Beetham 1991:6) that would mean: a citizen participating in a general election will to a greater extent than if not participating believe in the justifiability of the political authority. If the citizens justification of the authority stems from democratic principles and a belief that actual conditions sufficiently conforms to these principles, then we can in addition qualify the legitimacy as democratic legitimacy. The higher the turnout, the deeper and more widespread becomes the conviction of the citizens that political decision-making are justified and democratically justified. The higher the turnout, the higher is the expected level of democratic legitimacy. That’s my first assumption.

There is a least two reasons for the view that participation in general elections improves democratic legitimacy. First, participation in general elections can have an informative and educational effect. The procedure of choosing a political party and confirming the choice with the ballot, while following the political debate, can work as a fast course in political science and democratic theory, learning what democracy is about, the obligations and rights of democracy, its strengths and weaknesses. Given that the citizen are situated in a democratic culture and political order such information and deliberation are expected to increase the democratic legitimacy of the system, its democratic justifiability according to the citizens. Second, and to my view more important, participation in general elections improves democratic legitimacy because one can reasonably assume a spill-over between democratic praxis and conviction. If an individual behaves like a subject of the democratic decision-making procedure, that is he or she makes a political choice and walks to the ballot box, the democratic principles are expected to become more deeply embedded in that person. This would in turn create a deeper understanding for the majority principle, and accordingly that a minority has to accept also the decision it has voted against. Thus, participating in general elections are expected to yield democratic legitimacy to the political order.

To explore the relation between turnout in the European union and its member-states – and its supposed consequences in demo-

cratic legitimacy – we need to know what determines the voting behaviour. Why do people vote? Of course I can’t answer this question fully and seriously. During the last two days we have discussed different explanations, but taken together as one single model these explanations are fare too complex to use for a theoretical analysis about the relation between democratic legitimacy on the national and the European level. What to do then? My solution is a well-known in political science: when reality is too complex and you don’t know how to handle it, you turn to a ideal-type rational choice-model. Of course the model should be understood as an elaborated hypothesis and not a justified explanation. However, the interest of a rational-choice analysis is not only theoretical, as the idea of a citizen rationally calculating the political importance of voting or non-voting is an underlying assumption of a large number of arguments in public debate.2 My intention is to take these existing and influential arguments a step further.

A rationalistic answer to the question – why do people vote – would be that people vote because they want political influence, because they want to influence political decision-making or the scope of political decision-making. According to the same rationalistic hypothesis people don’t vote when they believe that voting makes too little political difference, that is: when the cost of voting is higher than its benefits, when for example voting takes too much time and energy compared to the gains of voting in terms of influence.

Before I go on to explain the relation between different democratic organisations sharing the same citizens I want to sum up the explanative assumptions made so far.


In this group we find among many others an argument pursued by

chairman Bengt Göranssons, saying that a low turnout in the European elections is expected as one of the most important questions of any political election, whether or not the government will stay, is never posed in the campaigns of the European elections (which is due to the constitutional limitations of the European parliament). If the elections were about something more important the argument implies that turnout would increase.

Diagram 1. A hypothesis

What the diagram says is first: that the citizens decision to vote or not to vote depends on the citizens desire for political influence at a minimal individual effort, and second: that the voting-act results in political legitimacy, which can be specified as democratic legitimacy because of its source in democratic principles and procedures.


If citizens vote because they want political influence the turnout in general elections is supposed to increase when adding more competencies and power to the organisation (given that it’s the elected body which controls the new powers, or rather: that citizens believe so). As there is a limited number of citizens and votes, each vote becomes more important when the organisation gains power, and the motivation for voting is thereby strengthened. This idea can be illustrated in a diagram.

Diagram 2. One democratic organisation, A

Note that we have not yet begun to compare different organisations. Both axis describe the same unit, organisation A. We try to capture only what is expected to happen in one single democratic organisation when the powers and competencies of the organisation are expanded. And what happens is that turnout is expected to increase.

If we stopped the analysis by here we would certainly conclude, that the solution to the problem of low turnout in the elections of the European parliament is to transfer more power to the Parliament, either from the member-states or from other political institutions at the international level, including the institutions of the European union. That is also a common recommendation for reducing the democratic problems of the union. For example one of the celebrities in this field, Joseph Weiler, professor of European law, has suggested that the European parliament should have power to take financial decisions – a power which is today exerci-

sed in the member-states – in order to make the citizens more interested in what the Parliament is doing (Weiler 1999:355). If the European parliament could decide about the money of the citizens then the citizens would be more interested to vote, according Weilers argument. We will however take the analysis a bit further and check if the conclusion is stable.

In what way could some other organisation, call it organisation B, differ from the first organisation A? In this ideal-type diagram the only thing that can vary is the inclination of the graph. So what could happen is that an organisation B has the following relation between turnout and the power of the organisation.

Diagram 3. One democratic organisation, B, which has more citizens OR more veto points than organisation A

Or we could think of an organisation where the variation in turnout is even smaller, as in this organisation C.

Diagram 4. One democratic organisation, C, which has more citizens AND more veto points than organisation A

How could we explain the differences between organisation A, B and C? If we want to stay in the rational-choice perspective there are at least two explanations for such differences. The first explanation is the size of the system, in terms of citizens. The second is the number of veto points in the decision-making procedure of the organisation. I comment shortly on each of them.

According to our rational-choice assumption the size of the system are expected to have a negative impact on turnout, all other things equal. The negative impact on turnout is expected because increasing the number of citizens implies a decreasing amount of political influence for each of them, and the incentive for voting is thereby weakened. Note the ”all other things equal” assumption. The system size is often thought to have a positive impact on system capacity, which could neutralise the negative impact of increa-

sing the number of citizens. All such very realistic consequences are excluded from this ideal-type diagram.

The second explanation to the differences in turnout between organisation A, B and C is the difference in the number of veto points in each organisation. The veto point is a stage in a decisionmaking procedure where the decision can be blocked. The number of veto points can vary between different organisations. For example, in a monist democracy – where antique Athens is probably the most extreme example as the political forum could not only legislate but also judge and directly execute their decisions – there is only one veto point. All that is needed for a social act to become a decision is the collective say at one moment and one place. This example could be contrasted with experiences from international negotiations where the final outcome has to be ratified in national parliaments before the proposal becomes a decision, that is: before the decision comes into power; in this example there are as many veto point as there are national parliaments, or even more, as there might be also constitutional and international courts. Now, if there are many veto points in an organisation rather than few, there will be less decisions taken, all other things equal, or at least: the decision-making will take more time and resources and thus be less efficient. The probability that a certain decision will be taken at a given time decreases, when the opportunities to block the decision increases. As the number of veto points decreases decision-efficiency, the number of veto points also are expected to decrease turnout in general elections; as participating in the elections of an efficient organisation is more interesting to citizen who seeks to maximise his or her political influence, than participation in the elections of an inefficient organisation. This holds for all kinds of veto points, but if some of the veto points are not controlled by elected bodies the problem is more severe, as the veto point then can act more independently.

The European union has more veto points than its memberstates, even if the number of veto points in the union vary between different policy areas. The decision-making process vary from regulating the common market, where there are at least three veto points – the Commission, the Parliament and the Council of mi-

nisters – to decisions on the competence of the European union, which takes ratification in the national parliaments. The turnout in the elections to the European parliament should reach a lower level because of the veto-points, compared to an institutional set-up with less veto-points.3

From a normative democratic perspective, the veto points of the European union are problematic. By adding veto points to a democratic decision-making procedure, the organisation gives more resistance to a majority of citizens supporting a political change. Instead of treating all possible alternatives alike, a decision-making procedure with many veto points favours status quo, which implies a bias of conservative ideology (thus violating the idea of a politically neutral decision-making procedure). This democratic problem holds for all kinds of veto points. But in the European union there are two further problems: (1) there is only one veto point which is directly elected (the other ones having their legitimacy mainly from non-democratic sources); (2) the formation of public opinion against and in support of the politics pursued by the European parliament, is hampered by linguistic, organisational and maybe cultural constraints. That turnout is lower in organisation C than in organisation A because its higher number of veto-points, could thus be re-phrased by saying that turnout is lower in organisation C than in organisation A because C is less democratic than A.

Let’s now turn to the comparison between turnout in different democratic units.

3 The relation is complex between turnout in the elections to the Euro-

pean parliament and the internal veto points of the Council of Ministers. If the Council of Ministers reaches a decision, its number of veto points could actually increase the influence of the European parliament, as the internal veto points weakens the Council of Ministers and thereby makes it easier for the Parliament to gain support for its position in a bargain with the Council. If the Council of Ministers however doesn’t reach a decision, the number of internal veto points will decrease the interest of the European parliament, as it can’t influence the situation.

Diagram 5. Two democratic organisations, A and B, in the political system (sharing the same citizens/voters)

This time we have organisation A on the vertical axis and organisation B on the horizontal axis. For the moment we have also included a simplification, a premise that says: every power and competence which is added to organisation B is taken from organisation A. Later on we will depart from this simplification. But for the sake of lucidity I start with a simple model. It’s not much to say about this diagram. As we have defined – in the premise – the power of organisation B as what is taken from organisation A, the graph should be inverted. The more power or competencies to or-

ganisation B from organisation A, the lower expected turnout in organisation A.

Now take a look at a diagram which includes the turnout in two democratic units in the same political system.

Diagram 6. Two democratic organisations with the same relevant characteristics, A:1 and A:2

Given that the organisations are alike in relevant aspects – number of citizens and number of veto points – we should expect a symmetrical trade off between the turnout in the two democratic units. What you gain in one unit you will lose in the other. Of course we should be aware that the overall system capacity is probably not the same when competencies and power are taken from one unit to the other in the same political system. But for still some time we stay with the ideal-type premise of the diagram.

We now come to the rather predictable comparison, between two more or less democratic units that differs from each other in aspects relevant for the level of turnout. Take for example a comparison between the organisations A and B, which we have previously identified, or the more provocative comparison between organisation A and C.

Diagram 7. Two more or less democratic organisations, A and B. B has either more citizens than A or more veto points

Diagram 8. Two more or less democratic organisations, A and C. C has both more citizens than A and more veto points

To make the situation even more obvious we can turn this diagram around and place organisation C on the horizontal axis instead of organisation A, while keeping the same relation between the two.

Diagram 9. Two more or less democratic organisations, A and C. C has both more citizens than A and more veto points

In this diagram we can easily see that it’s irrational to transfer any power from A to C or B, if you want to maximise turnout in general election. Yes, transferring power and competencies from organisation A to C or B will yield a higher turnout in C or B; but by doing this you will lose more turnout in A than you win in C or B.

As a small central state like for example Sweden is most resembling to A, and the large European union with several veto points is most resembling to C, we have now reached a quite different conclusion than the more simple one drawn by Weiler. His point was that an increase of power and competence of the European parliament would solve some of the democratic shortcomings of the union. At this moment we rather say that increasing the power of the European parliament would worsen the democratic

shortcomings of the union, though at the national level, as long as the competencies and powers transferred to the European parliament comes from national parliaments or bodies controlled by these.

But this conclusion is also premature. The premise – saying that every power and competence added to organisation A is taken from organisation C – is not true for all expansions of power and competencies in the European parliament. It’s thus time to problematize the premise and differ between policy areas where we believe that the premise is true and policy areas where we believe the premise to be false. I shall try to give a presumptive answer to this question.

What we are trying to identify is a situation where the power and competencies of one democratic unit can be expanded, without any negative effects on the other democratic unit. We thus look for a situation with the following structure.

Diagram 10. Two more or less democratic organisations, A and C. C has both more citizens than A and more veto points

Where can we find this? To use the term of Robert Dahl this structure would be find in areas where the power is alienated, and not delegated, from the democratic units (Dahl 1982:6, 47–53). If the power at present isn’t democratically controlled, then it can be added to another political level with no negative effects on turnout and democratic legitimacy. I will take three popular examples of this kind of power, which are commonly believed to need another political level than the nation-state for its effective implementation.

The first example is found in security policy. According to a popular realist thesis the European integration is taken forward by the necessity for nation-states to co-operate if they are to succeed in protecting their citizens from war. Put in another way: until the European integration, in the modern sense of this term, the secu-

rity policy of the European nation-states was a disaster of inefficiency. During less than seventy years Germany and France fighted against each other in three wars. So, if the security policy is transferred to the European level, that would not decrease turnout in national elections, as this policy couldn’t be effectively pursued on the national level anyway.

The problem here is that not all countries have had such a bad efficiency in security policy. Sweden hasn’t fought a war since 1809. You can accuse Sweden for having an immoral security policy during the second world war, and also for hiding its agenda after the war. But measured by its goal – to protect its citizens from aggressors – the policy has been efficient. It might then be difficult to convince a Swede that his or her state really has lost the possibility to pursue an efficient security policy.

The second example is found in environmental policy. Pollution’s of air and water can easily be taken from one nation to another; the same holds for global heating. According to a simple Prisoners Dilemma logic – that is where no state acts individually as all states need to act for solving a common problem – this could be a serious problem for the nation-states and might thus require an international level for effective governance. This is probably true. One problem however, is that environmental issues, though important and requiring international co-operation, might need a political level both above the European, and in some cases, below. For example, to rise the problem-solving-capacity of environmental policy in Sweden, its citizens need to co-operate with citizens in the Baltic states, Russia, Poland, Germany, England, US and China, but not to the same extent with the European citizens in Greece and Spain.

The third example is found in economic policy. The globalisation of capital and factors of production is often said to limit for example the tax-capacity of the nation-states. If capital, persons and factors of production easily can exit from a certain jurisdiction, no state will tax them higher than any other, which theoretically would yield a ”race to the bottom” of taxation on these taxbases. As the free movement of persons, capital, services and commodities is more intensive inside the European union than

outside, the union seems to be a feasible organisation for retaking some of the lost power to tax mobile tax bases. This should not affect the turnout in national elections, as long as the tax-power couldn’t be exercised nationally anyway. This time the problem is that such matters seem very hard for the member-states to agree on. Even if there is today social democrats in thirteen of fifteen member-state government, governments which are not by ideology opposed to public tax-capacity, the most radical decision up till now is the code of conduct, taken by the Council of Ministers in late 1997; and that is far less radical than the tax-competencies of mobile tax-bases which we are discussing now.

So, the conclusion from this three examples – security, environmental and economic policy – is that there are competencies and power which could be transferred to the European union, or competencies which has already been transferred, without decreasing turnout in the national elections, as the powers are not controlled by the national parliaments anyway. But we have also observed problems in each example, implying that the shortcomings of democratic legitimacy in the European union will not be satisfactorily solved by following the political implication of this analysis.

Finally then, my very simple answer to the initial question – about when it’s better or worse to divide a system in two levels, measured against a norm of democratic legitimacy – is that a division is better when the public that policies can be effectively pursued on one level but not at all on the other. And if this criterion is not fulfilled, we should keep decision-making power on the level with lowest possible number of citizens and veto points, if we want to maximise democratic legitimacy. To improve turnout it’s of course also directly elected bodies that should have decisionmaking power when competencies are transferred to a higher level.


Beetham, David (1991) The Legitimation of Power. Houndmills:

Macmillan Education LTD Dahl, Robert A. (1982) Dilemmas of Pluralist Democracy. New

Haven: Yale University Press Weiler, Joseph (1999) The Constitution of Europe. Cambridge:

Cambridge University Press

s å a r b e t a r

D e m o k r a t i u t r e d n i n g e n

Demokratiutredningen analyserar den svenska folkstyrelsens förutsättningar inför 2000-talet. Det sker i ljuset av bl.a. globaliseringen, EU-medlemskapet, förändringarna i medielandskapet, den nya IT-tekniken, förnyelsearbetet inom offentlig förvaltning och folkrörelsernas förändringar. Kommittén har också fått i uppdrag att förklara det sjunkande valdeltagandet samt stimulera det demokratiska engagemanget. Vid sekelskiftet skall utredningen presentera en sammanvägd analys i form av ett slutbetänkande.

Utredningen vill redan under arbetets gång bidra till att fördjupa demokratidebatten genom

en skriftserie

offentliga seminarier

en webbplats för debatt och information


Demokratiutredningen kommer fortlöpande under arbetets gång att publicera en skriftserie med essäer, debattartiklar, seminarieinlägg och forskarantologier.

Den omfattande demokratiforskning som bedrivs vid våra universitet och högskolor ställs på detta vis till förfogande för alla som är intresserade av demokratifrågor.

Information om titlar, utgivningsdag, priser m.m. finns på utredningens webbplats.


En av utredningens uppgifter är att stimulera det offentliga samtalet.

Till seminarierna bjuder vi in forskare och idédebattörer att bidra med sin kunskap och sina värderingar kring de frågor vi tycker är värdefulla att samtala kring.

Varje seminarium dokumenteras i form av en skrift.

www. demokratitorget.

är adressen till Demokratiutredningens webbplats. Här finns möjlighet att föra en fri debatt kring demokratin – demokratitorget är öppet för alla.

Ett antal krönikörer kommer att kommentera demokratiproblem och andra demokratidebatter.

Här kan Du också informera dig om och föra dialog kring utredningen. Direktiv i fulltext, arbetsplan och utredningens protokoll är dokument man kan hämta hem.

Här informerar vi också fortlöpande om aktuella skrifter och seminarier. Man kan dessutom ta del av debattinlägg vid seminarier samt beställa utredningens skrifter.

Länkar finns till andra delar av nätet där demokratifrågor dryftas på olika sätt.


Vi svarar gärna på frågor och lämnar ytterligare information om utredningens arbete. Adress och telefonnummer till utredningens kansli finns på nästa sida.

Demokratiutredningen analyserar den svenska folkstyrelsens förutsättningar inför 2000-talet. Det sker i ljuset av bl.a. globaliseringen, EU-medlemskapet, förändringarna i medielandskapet, den nya IT-tekniken, förnyelsearbetet inom offentlig förvaltning och folkrörelsernas förändringar. Kommittén har också fått i uppdrag att förklara det sjunkande valdeltagandet samt stimulera det demokratiska engagemanget. Vid sekelskiftet skall utredningen presentera en sammanvägd analys i form av ett slutbetänkande.

Utredningen vill redan under arbetets gång bidra till att fördjupa demokratidebatten genom

en skriftserie

offentliga seminarier

en webbplats för debatt och information


Demokratiutredningen kommer fortlöpande under arbetets gång att publicera en skriftserie med essäer, debattartiklar, seminarieinlägg och forskarantologier.

Den omfattande demokratiforskning som bedrivs vid våra universitet och högskolor ställs på detta vis till förfogande för alla som är intresserade av demokratifrågor.

Information om titlar, utgivningsdag, priser m.m. finns på utredningens webbplats.


En av utredningens uppgifter är att stimulera det offentliga samtalet.

Till seminarierna bjuder vi in forskare och idédebattörer att bidra med sin kunskap och sina värderingar kring de frågor vi tycker är värdefulla att samtala kring.

Varje seminarium dokumenteras i form av en skrift.

www. demokratitorget.

är adressen till Demokratiutredningens webbplats. Här finns möjlighet att föra en fri debatt kring demokratin – demokratitorget är öppet för alla.

Ett antal krönikörer kommer att kommentera demokratiproblem och andra demokratidebatter.

Här kan Du också informera dig om och föra dialog kring utredningen. Direktiv i fulltext, arbetsplan och utredningens protokoll är dokument man kan hämta hem.

Här informerar vi också fortlöpande om aktuella skrifter och seminarier. Man kan dessutom ta del av debattinlägg vid seminarier samt beställa utredningens skrifter.

Länkar finns till andra delar av nätet där demokratifrågor dryftas på olika sätt.


Vi svarar gärna på frågor och lämnar ytterligare information om utredningens arbete. Adress och telefonnummer till utredningens kansli finns på nästa sida.


Bengt Göransson

f.d. statsråd

tel: 08-453 41 04


Elisa Abascal Reyes (mp) tel: 08-519 560 06

Stefan Attefall (kd)

ledamot av riksdagen

tel: 08-786 54 46

Nils Fredrik Aurelius (m)

ledamot av riksdagen

tel: 08-786 46 85

Britt Bohlin (s)

ledamot av riksdagen

tel: 08-786 47 80

Mats Gerdau (m)


tel: 08-718 92 25

Per Gustafsson (c)


tel: 08-587 686 40

Gunnar Larsson (s)


tel 031-703 73 07

Eva Marcusdotter (s)


tel: 019-36 36 48

Olle Schmidt (fp)


tel: 040-34 10 64

Pernilla Zethraeus (v)


tel: 026-17 81 22


Lars Bryntesson


Wanja Lundby-Wedin

LO:s andra ordförande

Carl-Gunnar Peterson


Elisabeth Rynell


Marcus Storch



Huvudsekreterare Erik Amnå, fil dr tel: 019-30 30 67

Biträdande sekreterare Margareta Meyer tel: 08-405 49 16

Seminarieansvarig Anders Ljunggren tel: 08-405 16 54

Postadress: 103 33 Stockholm. Besöksadress: Regeringsgatan 30-32

Tel: 08-405 49 16. Fax: 08-411 24 67


Formgivning: Susan Nilsson Jupiter Reklam AB

Publications of the Commission of Democracy in Sweden

Editor Erik Amnå

1 . Demokratins räckvidd. Dokumentation från ett seminarium (SOU 1998:55)

2 . En god affär i Motala. Journalisternas avslöjanden och läsarnas etik (SOU 1998:63)

3. Att rösta med händerna. Om stormöten, folkomröstningar och direktdemokrati i Schweiz (SOU 1998:85)

4. Gör barn till medborgare! Om barn och demokrati under 1900-talet (SOU 1998:97)

5. Det unga medborgarskapet. Dokumentation från ett seminarium (SOU 1998:101)

6. Lekmannastyre i experternas tid. Dokumentation från ett seminarium (SOU 1998:102)

7. Demokrati på europeisk nivå? (SOU 1998:124)

8. Läsarna och demokratin. Ett brev till det läsande

Sverige (SOU 1998:134)

9. Lokala Demokratiexperiment


Exempel och analyser

(SOU 1998:155)

10. På Marginalen


En intervjubok från Socialtjänst-

utredningen (SOU 1998:161)

11. A l l a s frihet. Demokratin möter marknaden. Tre ronder (SOU 1998:164)

12.E U – e t t d e m o k r a t i p r o j e k t ? Dokumentation från

ett seminarium (SOU 1998:145)

13.In v a n d r a r s k a p o c h m e d b o r g a r s k a p .

Dokumentation från ett seminarium (SOU 1999:8)

14.A t t s l a k t a e t t f å r i G u d s n a m n . Om religionsfrihet

och demokrati (SOU 1999:9)

15. E t i k o c h d e m o k r a t i s k s t a t s k o n s t (SOU 1999:13)

16. E l e k t r o n i s k d e m o k r a t i (SOU 1999:12)

17.D e n s k yd d a d e p r o v i n s e n . En essä om

demokratins värde och värdighet (SOU 1999:22)

18.L o b b n i n g (SOU 1998:146)

19. D e m o k r a t i n i d e n o f f e n t l i g a s e k t o r n s f ö r ä n d r i n g . Dokumentation från ett seminarium (SOU 1999:40)

20.B e m ä k t i g a i n d i v i d e r n a . Om domstolarna, lagen och

de individuella rättigheterna i Sverige (SOU 1998:103)

21. B ö r d e m o k r a t i n a v n a t i o n a l i s e r a s ? (SOU 1999:11)

22. G l o b a l i s e r i n g e n o c h d e m o k r a t i n . Dokumentation från ett seminarium (SOU 1999:56)

23. L ö s e r j u r i d i k e n d e m o k r a t i n s p r o b l e m ? (SOU 1999:58)

24. R e p r e s e n t a t i v d e m o k r a t i (SOU 1999:64)

25. D e m o k r a t i n o c h d e t g e m e n s a m m a b ä s t a (SOU 1999:74)

26. D e m o k r a t i o p i n i o n e r (SOU 1999:80)

27. O l yd i g a m e d b o r g a r e ( SOU 1999:101)

28. Demokratins förgörare (SOU 1999:10)

29. Civilsamhället som demokratins arena. Några kritiska reflektioner (SOU 1999:112)

30. Demokrati på remiss (SOU 1999:144

31. Vad hände med Sveriges ekonomi efter 1970? (SOU 1999:150

32. Citizen Participation in European Politics (SOU 1999:151)