Resumé

The aim of this article is to analyze the explanations for the increased number of referendums in Western European states. Bogdanor has explained this by the unfreezing of political alignments institutions. After classifying referendums in 18 Western European states, it will be observed that especially the number of referendums introduced by citizens has increased wherever they are constitutionally provided, that is, in Italy and in Switzerland. Referendums in these two countries contribute most of the overall increase in the number of referendums in Western Europe. Moreover, the number of referendums concerning European integration have increased, also in states that do not have the referendum as an integral part of their political system. When it comes to the unfreezing of institutions, it must be pointed out that after the 1970s there have been no constitutional reforms which would have significantly increased the use of referendums.

Introduction

Bogdanor characterizes the developments in the use of referendums in
advanced democracies as follows (1994, 91):

It seems then as if some general cause may be at work in the advanced democracies that leads to increasing use of the referendum. That common cause is likely to be the unfreezing of the political alignments and institutions formed after the achievement of universal suffrage in the first two decades of the twentieth century.

Bogdanor refers to the famous freezing hypothesis by Lipset & Rokkan (1967). According to the hypothesis, the current party systems in Western European democracies reflect the socio-economical structures of the early 20th century. By the unfreezing of alignments, Bogdanor (1994, 94) refers to the tendency that 'the importance of social-structural determinants of voting behavior, whether religion or class, has declined.' This means that voters no longer necessarily 'belong' to one party according to their social rank, and that voters may also have become more critical of the traditional

---

* Maija Setälä, Department of Political Science, 20014-University of Turku. Finland. E-mail: maiset@utu.fi

parties as their representatives. By the unfreezing of institutions, Bogdanor refers to the fact that the representative institutions which were established in the late 19th and early 20th century have been challenged by the demands of more participatory forms of democracy since the 19605.

The purpose of this article is, first, to refine Bogdanor's general explanation the increased number of referendums and to make a more careful analysis of the causal relationship between the unfreezing of alignments and the increase of referendums. For this purpose, I analyze what kinds of referendums have increased, and discuss the mechanisms through which the unfreezing of political alignments has increased referendums. Second, another factor that has increased the use of referendums, that is, the European integration process is scrutinized. Third, I examine whether the pressures on representative institutions have materialized as constitutional reforms which significantly increase the numbers of referendums.

A Classification of Referendums

The questions raised in this article cannot be answered adequately without knowing what kinds of referendums are discussed. There are several classifications the referendums put forward for example by Smith (1976), Suksi (1993), and Gallagher & Uleri (1996). In order to analyze the reasons for the increase of referendums, it is necessary to focus on how referendums are introduced.

Referendums may be either mandatory or non-mandatory (facultative). Mandatory referendums are constitutionally required upon certain issues. Non-mandatory referendums are triggered only at certain political actors' request. Furthermore, using Suksi's terminology, non-mandatory referendums be either passive, when they are introduced by the representatives, active, when they are introduced by a certain number of citizens. In addition to the initiator of the referendum, the source of the legislative proposal on which the referendum is held needs to be taken into account. The terminology used here originates from Uleri (1996a, 10-12), who puts forward a distinction between law promoting and law controlling referendums. law promoting referendums the author of the law to be voted upon also introduces the referendum. In law controlling referendums, the author of the law proposal does not introduce the referendum. Table 1 summarizes the classification of referendums basedon two criteria¹:

Mandatory referendums are law controlling per definition, because they are constitutionally required on certain issues. Non-mandatory law controllingreferendums be introduced by a certain number of citizens, by an actor in the representative government, or by local or regional governments.Moreover, referendums are held upon legislative proposals

which have been accepted by the parliamentary majority, but which have
not yet been enacted. Abrogative referendums are held on laws that are
already in force.

According to Uleri (1996a, 12), the distinction between optional and ad hoc referendums is that optional referendums are constitutionally regulated and ad hoc referendums are not." This distinction is insignificant in practice when the parliamentary majority makes the decision to have a referendum. There are also a few cases in which other actors in the representative government besides the parliamentary majority introduce optional referendums. example, in France the power to introduce a referendum is shared by the president and the parliament. A popular initiative refers to a law promoting referendum introduced by a certain number of citizens. The initiative is made by collecting signatures to support a law proposal which is drafted by a citizens' committee.

Different forms of referendums have rather different consequences with respect to the influence exerted by different political actors in the decision making process. In general, law controlling referendums function as a retrospective check on parliamentary decisions, and they are adopted to protect the legislative status quo. Mandatory referendums have functional similarities with multi-cameral decision making. Abrogative and rejective referendums are particularly aimed at protecting the interests of minorities, who may block the decision making process by demanding a referendum. Popular initiatives allow citizens to set the political agenda by raising issues to be voted upon in referendums. Because ad hoc and optional referendums are usually introduced by the parliamentary majority, the power to introduce kinds of referendums is typically in the hands of the governmental or coalition.

Mandatory referendums are triggered 'automatically' on certain types of issues, normally on constitutional amendments, although sometimes the question of the constitutional necessity of a referendum on a certain issue has become a matter of political dispute. The initiation of popular initiativesand law controlling referendums may normally be explained by 'policy maximizing' motivations (cf. Hug 1995). The initiators of these kinds of referendums want to promote a certain law proposal, or

to defend the legislative status quo against a proposal that they find unfavorable. At the first sight, the interests behind ad hoc or optional referendumscannot regarded as 'policy maximizing,' because the parliamentarymajority to introduce a referendum is normally also sufficient to pass the proposal submitted to a referendum. Exceptions to this rule are referendums introduced by minority governments and referendums held on issues on which a qualified majority is required in the legislature. Other possible motivations behind ad hoc and optional referendums are discussed later.

Institutions in Western Europe

In general, it may be assumed that parties most likely to be in the governmental office are particularly cautious to adopt law controlling and especially popularly introduced referendums, because their use is largely beyond government control, and they may be targeted against government policies. This section describes the constitutional provisions for referendums 18 Western European states, and briefly reviews how they have been adopted. In Table 2, 18 Western European states are classified according the type of referendum provided in their constitutions. Because ad hoc referendums are not constitutionally regulated, all states that have experienced these kinds of referendums since 1940 are included in the Table. The year the constitutional provisions for referendums were adopted are given in brackets (cp. Gallagher 1996b, 226-29):

Of the constitutional provisions for mandatory constitutional referendums, those of Denmark, Ireland and Switzerland have been put into practice. In these states, mandatory referendums are required on all constitutional whereas in Spain and in Austria mandatory referendums required only on fundamental changes to the constitution, and in Iceland only on certain types of constitutional amendments.

The Danish case illustrates how the referendum has been adopted as an institution that limits the powers of parliamentary majorities. The mandatory referendum was first adopted in Denmark in 1915 to compensate the abolition of the requirement that constitutional changes should be passed in two subsequent parliaments. In 1953, different forms of law controlling referendums were adopted in conjunction with the abolition of the upper chamber of the parliament, the Landsting. This was a result of a compromise between those parties supporting the idea of popular sovereignty and parties concerned about the interests of minorities (Suksi 1993, 183-86; Svensson 1996, 35-36).

The Constitution of the Irish Free State (1922) included a provision for
a mandatory referendum and a popular initiative. The Constitution of the

Irish Free State reflected the ideas of popular sovereignty and republicanism.Thecurrent for mandatory referendums were introduced in the Constitution adopted in 1937 (Chubb 1983; Bogdanor 1994, 78-79; Gallagher 1996a, 86-106). The Swiss mandatory constitutional referendum dates back to the year of the foundation of the modern Swiss federation in 1848. Since 1977, a mandatory referendum has also been required on issues concerning membership in collective security organizations and supranationalcommunities(Trechsel Kriesi 1996). The direct-democratic traditions of Swiss cantons, the ideology of popular sovereignty of the French Revolution, and the willingness of the cantons to restrict the powers

of the federal government have been seen as reasons for the adoption of
different forms of referendums in the Swiss Federal Constitution.

Of the constitutional provisions for rejective and abrogative referendums, only those in Denmark, Italy and Switzerland have been put into practice. Since the 1953 constitutional reform, it has been possible for one third of the members of the Danish Folketing to introduce a rejective referendum. In Italy, the law on implementing the constitutional provision for abrogative referendum (dating back to 1948) was adopted in 1970. Abrogative referendums be introduced by 500,000 voters or by rive regional councils. The implementation of the constitutional provision for abrogative referendum a condition put forward by the Christian Democratic Party to its governmental coalition partners to approve the law on divorce. In other words, the abrogative referendum was adopted as part of a compromise between the governmental parties. The first abrogative referendum in 1974 was, consequently, on divorce (Uleri 1996b, 110-11). The Italian abrogative referendum is a rather special case of law controlling referendums, because there is no limit on the length of time between implementation of the law to be voted upon and the referendum. Therefore, the Italian abrogative referendum had similar impacts on the political agenda as the popular initiative (ibid., 109).

In Switzerland, since 1874, the number of voters required to demand rejective referendums on normal laws is 50,000 (until 1977, it was 30,000). Since 1949, the same number of people have had a right to demand abrogative on laws declared urgent. This constitutional reform was based on a popular initiative. Switzerland is the only Western European state with constitutional provisions for popular initiatives at the national level. Since the foundation of the Swiss Confederation in 1848, Swiss citizens have had the right to take initiative to a total revision of the constitution, and since 1891, to partial revisions. The number of signatures currently required for an initiative is 100,000.³

Ad hoc and optional referendums remain the most widespread types of the referendum in Western European states. On many occasions, ad hoc and optional referendums have been used as an external arbitrator in case of internal splits in parties and governmental coalitions. The referendum on the supplementary pension plans in Sweden in 1957 is an example of how the referendum can be used to solve a parliamentary deadlock. Maintaining the unity of the party was the motivation, for example, behind the British referendum on staying in the EEC in 1975 promoted by the Labour party (Bogdanor 1981, 90-93; Balsom 1996). The referendum may be used to shape the electoral agenda in order to avoid unpopular issue positions. A referendum has been used as a means to separate issues from the electoral agenda, for example, in Sweden in conjunction with the referendum on nuclear power in 1980.⁴

Increased Number of Referendums?

In relation to the question of whether the unfreezing of political alignments causes referendums, non-mandatory referendums are of special interest, because their introduction depends on certain political actors' interests. The link between active referendums and the unfreezing of alignments is relatively an increased number of active referendums can be expected if citizens become increasingly detached from the traditional parties. Also the number of passive abrogative and rejective referendums may rise if governmental parties increasingly fail to represent the opinions of the majority of voters on a salient issues. Finally, the use of ad hoc and optional referendums may increase as a means to resolve issues that cross the prevalent political dimensions.

Table 3 gives the numbers of nationwide referendums in Western European since 1940. It must be emphasized that this table does not include referendums held at the sub-state level, such as the ones held on devolution in Britain in 1979 and in 1997, or referendums held at the local level.

Switzerland is by far the most frequent user of referendums. Also Denmark, Ireland and Italy have experienced several referendums. France has been the most active user of optional referendums, especially during President de Gaulle's rule in the 19605. Figure 1 describes the average number of referendums per year between 1940 and 1998. At first sight, the overall number of referendums in Western European states shows a clearly increasing trend, from an annual average of 2.6 referendums in the 1940s to 15.6 in the 19905. Moreover, the number of states that have had referendums also increased slightly. This may explain why referendums have become of public interest also in states in which they are not an integral part of the political system.

However, if referendums in Switzerland and Italy are counted separately, the picture changes somewhat. Switzerland and Italy are the only states where active referendums are constitutionally provided. The number of referendums per decade in Switzerland has constantly been higher than in all other states combined. In Italy, the adoption of the active abrogative referendum in 1970 prompted a significant increase in the number of referendums. in the 1980s and 19905, Italian referendums constituted more than half of the nationwide referendums held in Western Europe outside Switzerland. Figure 2 illustrates the numbers of referendums in Switzerland, Italy and in other 16 states separately. Table 4 presents the numbers on which Figures 1 and 2 are based.

Figure 2 indicates that active referendums are increasingly used where they are constitutionally provided. Of the 47 referendums held in post-war Italy, 45 have been abrogative. Of these, 43 referendums have been active, that is, introduced by collecting citizens' signatures. Regional councils have introducedtwo Moreover, the number of Italian abrogative referendumshas

endumshasincreased from three in the 1970s to 31 between 1990 and 1998.
According to Uleri (1996b, 120), the rapid increase of referendums was
unexpected when the law on abrogative referendums was adopted in 1970.

In Switzerland, about 30 percent of all referendums between 1940 and 1998 were based on popular initiatives. In addition to these, a significant number of initiatives have been withdrawn in negotiations with the Federal Government and therefore not voted upon in a referendum. Popularly introduced abrogative and rejective referendums constitute roughly 25 percent of all Swiss referendums held between 1940 and 1998; the rest were mandatory. The relative frequencies of different types of referendums have remained relatively stable over time. This means that also the number of mandatory referendums has been growing. The increase of mandatory referendums has been explained by the expanding scope of the federal legislation (Trechsel & Kriesi 1996, 191).

According to some scholars, a large part of popular initiatives in Switzerland have been introduced by political outsiders and by new social groups. Swiss abrogative and rejective referendums, on the other hånd, have been more of an instrument of political insiders (Kobach 1993). In Italy, political entrepreneurs and protest parties and groups (e.g., The Radical Party, far leftist groups, environmental groups) have used the abrogative referendum (Uleri 1996b). These midings may be interpreted as an indication an increasing sentiment felt by political activists and citizens that the representative system and the traditional parties do not adequately represent the interests and issues they find important. This is in accordance with Bogdanor's explanation of the unfreezing of political alignments as a rcason for the incrcasecl numbcr of referendums.

However, there seems to be some increase in the number of referendums even among users of other forms of referendums, especially in the 19905. It is easy to point out one explanation for this development, that is, the European integration process. The integration issue has become an important of referendums. There have been altogether 19 referendums on this issue in all 18 Western European states. There were five referendums related to these issues in the 19705, three in the 1980s, and eleven in the period 1990-98. Of the nationwide referendums held in other Western European states than Switzerland and Italy in 1990-98, 10 out of 19 concerned integration issue.

In the 19705, there were five referendums on the enlargement of the European Community (Britain, Denmark, France, Ireland, Norway). In the 1980s, there were three referendums on the 'deepening' of the Community: and Ireland held referendums on the Single European Act, and the Italians voted on drafting the Treaty of European Union. During the 19905, Austria, Finland, Norway and Sweden have held referendums membership in the European Union. The Swiss voted on membership in the European Economic Area in 1992. Denmark, France and Ireland held referendums on the ratification of the Maastricht Treaty in 1992, and in 1993 Denmark held a referendum on the Edinburgh Agreement after the rejection of the Maastricht Treaty. In 1998, Denmark and Ireland held referendums on the Treaty of Amsterdam.

In Denmark, Ireland and Switzerland, issues related to participation in the integration process have been submitted to mandatory referendums.⁵ The explanation for the increasing use of ad hoc and optional referendums on the integration issue may be found in the need to legitimize decisions related to the integration process. However, referendums have been used for other purposes than legitimation. Especially, it seems that referendums have been used by parties as a strategy to deal with divisions caused by the integration issue. The integration issue has divided parties and governmentalcoalitions, example in Norway in 1972 and 1994, in Britain in

1975 and in Sweden and in Finland in 1994. The observation that ad hoc and optional referendums have been held on the integration issue, as well as other issues which divide traditional parties or coalitions, is in accordance with Bogdanor's explanation that the unfreezing of alignments increases the use of referendums.

Referendums on integration issues have, by no means, been an unproblematic of resolving these issues: Four times (Norway 1972 and 1994, Denmark 1992, Switzerland 1992 on EEA agreement), voters have rejected proposals to participate in integration put forward by the governments, in France in 1992, the 'yes' side won by a very small margin. Referendums on the integration issue have revealed the differences between the public opinion and the opinions of the political elites. It has been argued that the referendum results have had more to do with the unpopularity of the governments rather than the unpopularity of the integration issue (Franklin et al. 1994). Although it makes a difference how and by whom the issue has been presented to the voters, this view seems rather extreme considering the evidence provided by opinion polls on the continuously critical public opinion on integration in certain countries. The critical public opinion may also have made the political elites more reluctant to promote a referendum as a means to resolve these issues.

Conclusions

Bogdanor explains the increase of referendums by a 'general cause,' that is, the unfreezing of alignments and institutions. I argue that unfreezing of political alignments has increased active forms of referendums where they are constitutionally provided. Another obvious explanation for the overall increase of referendums is the growing number of referendums on issues related to European integration. Ad hoc and optional referendums on European integration have been experienced especially in such countries where the issue has been salient and divisive, which is in accordance with Bogdanor's explanation.

Although the unfreezing of traditional political alignments and the rise of new political dimensions may have caused public demands for more referendums, these demands have not been materialized to the full extent. Bogdanor admits (1994, 96): 'It is understandable, however, that where the referendum remains under the control of the political class, that class does not have much taste of it.' In addition to causing delays and being a quite expensive means of resolving issues, a referendum may also bring about rather unexpected results. These factors explain why the increase in the number of referendums has remained relatively modest outside those states where they are introduced by citizens.

There may be some objections to the notion of the unfreezing of political institutions. Although radical democrats and various theorists of democracy expressed their support for law controlling and active forms of referendums (cf. Barbér 1984), it is, however, unlikely that constitutional provisions for these kinds of referendums will be widely adopted. One obvious reason is that these types of referendums are beyond the control of governments and they may be directed against the policies pursued by governments. Therefore, it may be expected that they confront opposition by the parties most likely to be in office. This argument may be supported, for example, by the experience of Sweden, where the proposals made by constitutional reform committees since 1920s to expand the use of law controlling referendums have been watered down by the main parties, and where the reforms made in 1979 were eventually minor (see Suksi 1993, 212-19; Ruin 1996).

The last two significant constitutional reforms to increase the use of referendums outside Switzerland (1977) have taken place in Denmark (1953) and in Italy (1970). The Danish reform extended the use of referendums protect the interests of minorities as a part of a large constitutional The adoption of the Italian abrogative referendum was an element of a compromise between parties on a salient issue, and its consequences were unexpected at the time. It seems that the main parties are generally cautious to adopt such unpredictable and uncontrollable elements in national politics as popularly introduced and law controlling referendums. The Italian experience of abrogative referendums and also more sporadic events like the Danish referendum on the Maastricht Treaty are good examples of this unpredictability.

---

NOTES

1. In addition to these categories, referendums may also be classified as binding or consultative. referendums either have immediate normative consequences or the parliament is constitutionally obliged to follow the outcome of the referendum. Mandatory and popular initiative-based referendums are typically legally binding, just as rejective and abrogative referendums. Optional referendums may be binding if it is mentioned in the constitution. Ad hoc referendums are usually legally consultative, although they may be considered de facto binding.

2. The distinction between those referendums which are constitutionally regulated and those which are not is relevant only with respect to referendums introduced by the parliamentary majority, because all other forms of referendums are constitutionally regulated.

3. In the Netherlands, a constitutional reform committee has suggested adopting a rejective introduced by a certain number of voters. The institution suggested in The Netherlands most resembles the Swiss institution of rejective referendums, although the suggested thresholds for making the initiative is considerably higher than in Switzerland. It is questionable whether the proposal will be adopted (Andeweg 1997).

---

3. In the Netherlands, a constitutional reform committee has suggested adopting a rejective introduced by a certain number of voters. The institution suggested in The Netherlands most resembles the Swiss institution of rejective referendums, although the suggested thresholds for making the initiative is considerably higher than in Switzerland. It is questionable whether the proposal will be adopted (Andeweg 1997).

4. Moreover, in the Swedish referendum on European Union membership in 1994, the Social Democratic Party found it important not to have the referendum simultaneously with the parliamentary election in September, but rather two months later in November, because the party leadership thought that simultaneous campaigns could lead to an electoral defeat (see Gilljam 1996, 19).

5. In Denmark, most referendums on the integration issue (apart from 1986 and 1993) have been held under the constitutional provision for referendums on issues concerning transfer of national power to international organizations. The referendum is needed if a proposal on such an issue fails to get the support of five sixths of the members of Folketinget. In Ireland, all four referendums on the integration issue (held in 1972, 1987, 1992 and 1998) have been held under the provision for the mandatory constitutional The Irish Supreme Court decided in 1986 that the Single European Act implied a constitutional change, and therefore also the subsequent steps taken to 'deepen' the European integration process have been submitted to mandatory constitutional referendums (Gallagher 1996a).

ACKNOWLEDGMENTS

I would like to thank Hannu Nurmi and the anonymous referees for their comments on this
article.

REFERENCES

Andeweg, R. B. 1997. 'Institutional Reform in Dutch Politics: Elected Prime Minister,
Personalized PR, and Popular Veto in Comparative Perspective,' Acta Politico 32/3.
227-57.

Balsom, D. 1996. 'The United Kingdom: constitutional pragmatism and the adoption of the
referendum,' in Gallagher, M. & Uleri, P., eds.. The Referendum Experience in Europe.
Basingstoke: Macmillan.

Barber, B. R. 1984. Strong Democracy. Participatory Politics for a New Age. Berkeley-
University of California Press.

Bogdanor, V. 1981. The People and the Party System. The Referendum and Electoral Reform
in British Politics. Cambridge: Cambridge University Press.

Bogdanor, V. 1994. 'Western Europe,' in Butler, D. & Ranney, A., eds., Referendums around
the World. Basingstoke: Macmillan.

Chubb, B. (ed., assisted by O'Dea, G.). 1983. A Source Book of Irish Government. 2nd edition.
Dublin: Institute of Public Administration.

Franklin, M., van der Eijk. C. & Marsh, M. 1994. 'Referendum outcomes and Trust in
Government: Public Support for Europe in the wake of Maastricht,' Western European
Politics 3/18, 101-18.

Gallagher, M. 1996a. 'Ireland: the referendum as a conservative device?,' in Gallagher. M. &
Uleri, P., eds., The Referendum Experience in Europe. Basingstoke: Macmillan.

Gallagher, M. 1996b. 'Conclusion,' in Gallagher. M. & Uleri, P. eds.. The Referendum
Experience in Europe. Basingstoke: Macmillan.

Gallagher, M. & Uleri, P. Eds. 1996. The Referendum Experience in Europe. Basingstoke:
Macmillan.

Gilljam, M. 1996. 'Den direkta demokratin,' in Gilljam, M. & Holmberg. S., eds.. Ett knappt
ja till EU. Stockholm: Norstedts Juridik.

Hug, S. 1995. 'What Causes Referendums? Uncertainty? Preferences? Institutions?.' Paper.
Annual Meeting of APSA, Chicago.

Kobach, K. 1993. 'Recent Developments in Swiss Direct Democracy,' Electoral Studies. 4
12,342-65.

Lipset. S. M. & Rokkan, S. 1967. Party Systems and Voter Alignments: Cross-National
Perspectives. New York: Free Press.

Research and Documentation Centre of Direct Democracy at the University of Geneva
(www pages: http://c2d.unige.ch, January 31, 1999.).

Ruin, O. 1996. 'Sweden: the referendum as an instrument for defusing political issues,' in
GalUishcr, M. & Uleri, P., eds., The Referendum Experience in Europe. Basingstoke:
Macmillan.

Setälä, M. 1999. Referendums and Democratic Government. Basingstoke: Macmillan.

Suksi, M. 1993. Bringing in the People. Dordrecht: Martinus Nijhof Publishers.

Svensson, P. 1996. 'Denmark: the referendum as minority protection,' in Gallagher, M. &
Uleri, P., eds., The Referendum Experience in Europe. Basingstoke: Macmillan.

Trechsel, A. & Kriesi, H. 1996. 'Switzerland: the referendum and initiative as a centerpiece of
the political system,' in Gallagher, M. & Uleri, P., eds., The Referendum Experience in
Europe. Basingstoke: Macmillan.

Uleri, P. 1996a. 'Introduction,' in Gallagher, M. & Uleri, P., eds., The Referendum Experience
in Europe. Basingstoke: Macmillan.

Uleri, P. 1996b. 'Italy: referendum and initiatives from the origins to the crisis of a democratic
regime,' in Gallagher, M. & Uleri, P., eds., The Referendum Experience in Europe.
Basingstoke: Macmillan.