From The Referendum Experience in Europe, Eds. Michael Gallagher and Pier Vincenzo Uleri, New York: St. Martins Press 1996, 33-51. Notes and references provided by the original source have been omitted.

Denmark: the Referendum as Minority Protection

By Palle Svensson, Professor dr scient pol, Aarhus University

Since the mid-1980s, three referendums have taken place in Denmark on European integration and, in addition, many demands and proposals have been put forward for further referendums, from both the government and the opposition. Moreover, there is broad political acceptance that any future change in Denmark's relationship with the European Union will have to be put before the voters for final approval or rejection.

Accordingly, referendums seem increasingly to be regarded as an appropriate measure for solving political conflicts in Denmark, despite the fundamentally representative character of the political system. Even if the Danes take high pride in their political system and think of it as a "democracy", the democratic idea is not realized any more completely in this country than elsewhere. Denmark is a small country with five million inhabitants and four million voters, but this does not mean that the political system is more perfect than in other countries, or that it is more direct than representative.

The Constitutional Background

The Danish political regime is not a realization of any specific democratic theory. It should rather be seen as a result of a long process, where various ideas and group interests have confronted one another under changing social and political conditions. This process has led to continued democratization of the political regime, despite occasional setbacks for the democratic forces. Unlike its counterparts in Sweden and Finland, the Danish constitution does not directly express the principle of the sovereignty of the people. Even so, it is reasonable to say that this principle forms the ideological basis of the regime.

The principle of the sovereignty of the people is realized by means of a representative form of government. The citizens directly elect the Danish parliament (the Folketing) and thereby have an influence on the composition of the government and thus, indirectly, on policy making. The combination of the two main principles in Danish democracy - the principle of popular sovereignty and the principle of representative government - is expressed in universal suffrage; in proportional representation ensuring all political opinions an accurate representation in the Folketing; and in the parliamentary system making the government responsible to the majority in the Folketing.

The democratization process has, in addition, allowed the citizens some opportunity for direct influence on policy making by means of referendum. Elements of direct democracy have come rather late in the process of democratization in Denmark and have been met by severe opposition from defenders of representative democracy.

The 1915 constitution introduced the mandatory referendum as a part of the procedure for constitutional amendments. Previously such amendments could be made if a bill was passed by parliament and, following an election, passed again by the new parliament. Under the 1915 constitution an amendment proposal - when passed by parliament and afterwards by the newly elected parliament - had to be submitted to the voters for approval or rejection. Approval of the amendment required a majority among the participating voters, and at least 45 per cent of the whole electorate (in the 1953 constitution, this was reduced to 40 per cent).

This referendum gives effect to both the principle of popular sovereignty and the principle of minority protection. It is in accordance with the principle of popular sovereignty that the constitution must be based on the popular will and that amendments must be put direct to the electorate for decision by majority. On the other hand, it is in accordance with the principle of minority protection that it is very difficult to change the constitution, since this requires not just a majority among those voting but also the explicit approval of a substantial proportion of the electorate.

The 1953 constitution introduced the optional or facultative referendum. Its history goes back to the beginning of the century, and as early as the 1930s a degree of agreement had been reached between the major political parties about the fundamental outlines. Under Article 42 of the constitution, one third of the members of the Folketing can demand that a bill passed by parliament be submitted to the voters for either approval or rejection. Rejection of the bill requires a negative majority that comprises at least 30 per, cent of the electorate.

An analysis of the genesis and provisions of this referendum makes it quite clear that its basis is neither the principle of popular sovereignty nor the aim of giving the voters direct influence on decision-making. Rather, this rejective referendum must bee seen as a protection of the minority in the Folketing against the majority. Proposals for a rejective law referendum entered the constitutional negotiations late in the 1930s. The constitutional motion from the Social Democratic-Radikal Liberal government contained no mention of referendums. Among other things, the Liberals and the Conservatives criticized the bill for abolishing the Landsting (the upper chamber) without combining this with some kind of protection of the interests of the minority. 

During negotiations, the government and the Conservatives reached a compromise, which included a provision for an optional law referendum. This compromise was a fundamental break with all previous ideas and proposals about referendums in Denmark. The principle of popular sovereignty was downplayed in favor of the protection of the minority in parliament, and direct democracy was disregarded in favor of representative democracy. The proposals would have given the people very limited opportunities to demand a referendum compared with some previous proposals. Representative democracy was clearly given the upper band, since the decision of parliament on a law was to prevail unless there was a popular majority, comprising 35 per cent of the electorate, against it. Whereas abstention by voters in a referendum on a constitutional amendment has the effect of expressing opposition to a proposed amendment of the constitution, abstention in a referendum on a law is an expression of faith in and acceptance of the decisions of the elected representatives.

These proposals were not implemented, because the constitutional amendments proposed by the Social Democrats, Radical Liberals and Conservatives were narrowly defeated in the referendum in May 1939 - they were approved by 44.46 per cent of the electorate, compared with the required 45 per cent - and new discussions on constitutional changes did not start until the end of the war and the German occupation. In 1953 a comprehensive proposal for a new constitution was supported by the four old parties (Social Democrats, Radical Liberals, Conservatives and Liberals) and the Single Tax Party (the Henry George-inspired Retsforbundet). This proposal included the provision for an optional rejective law referendum that forms Article 42 in the present constitution.

This provision is elaborated in much the same way as the one defeated in 1939. However, the last remnant of any possible right for the voters to play a part in demanding a referendum was removed, as under this Article one-third of the members of the parliament - the single-chamber Folketing - can demand a referendum. Participation by the voters is neither necessary nor possible. The conditions needed to reject a bill - a majority of votes, comprising 30 per cent of the electorate - represent a compromise between the 35 per cent proposed in 1939 and the 25 per cent proposed by the Liberals in the 1930s.

To summarize, the evolution of thinking on the optional law referendum demonstrates a movement away from the principle of sovereignty of the people and direct democracy towards minority protection and representative democracy. The option chosen in 1953 was the one giving the majority of the people the smallest opportunity of influencing legislative decisions among the proposals discussed.

Apart from the rejective law referendum, the 1953 constitution also introduced a mandatory referendum on the voting age and a mandatory referendum on delegation of Danish sovereignty to international authorities if a bill on this subject is not approved with a majority of five-sixths of the members of parliament. The requirements for rejection laid out in Article 42 also apply here; that is, there must be a negative majonty comprising at least 30 per cent of the electorate in order to defeat the passed bill.

Both these referendum options make it easier to implement changes that had hitherto entailed constitutional amendments. This is most marked with regard to the voting age, which was previously defined in the constitution. With regard to delegations of sovereignty, the requirement of a five-sixths majority is rather drastic, but the degree of popular support needed for such a bill in a referendum is lower than that needed to approve a constitutional amendment.

These provisions represent a step away from the sovereignty of the people and direct democracy towards representative democracy. In both cases, it has become easier for the Folketing to make decisions that previously required an explicit and strong endorsement from the voters. Now it is required on these issues only that a majority of the voters of a considerable size (30 per cent of the electorate) do not reject the decision of the Folketing. This clearly favors representative democracy at the expense of direct democracy.

Finally, it is generally assumed that there exists the possibility of ad hoc and advisory referendums - a possibility that is not regulated or provided for in the written constitution. After a proposal on an ad hoc referendum was not included in the final bill on constitutional amendments that was accepted in 1915, there was some debate as to whether ad hoc referendums would be precluded in the future. Shortly after this, however, an ad hoc and advisory referendum for the sale of the Danish West Indian Islands (now the US Virgin Islands) was carried through in 1916. The general view among constitutional experts is that nothing prevents the holding of referendums outside the framework of the constitution. The result of such referendums may serve as a guidance for competent governmental institutions, though they cannot legally restrict the freedom of decision-making of theses institutions.

There are no legal or constitutional provisions concerning the funding of opposing sides in a referendum campaign, but as part of a general commitment to political fairness it is broadly accepted that public funds should be made available to both sides, and pragmatic solutions have been found. The details vary from campaign to campaign, but in general quite substantial sums have been provided; these have been allocated by an independent committee, and a nongovernmental organization (the Danish Council for Adult Education) has carried out the administration. As of mid-1995, funds (currently amounting to 15 million kroner, approximately £1.7m sterling) for a future referendum on further European integration have been equally divided between established organizations (political parties and popular movements for and against more integration) and individual applications for meetings, publications, and so on. Applications are considered on their merits and not necessarily with a view to achieving absolute equality in the allocation to the two sides.

To sum up, this analysis of the adoption of referendums in the Danish constitutional system makes it clear that the principle of popular sovereignty, as expressed in various forms of direct democracy, has penetrated the constitutional framework to only a limited extent. Only when it comes to constitutional amendments do the people as a whole have the possibility of direct political decision-making, and even this is qualified by the stipulation that a favorable majority must comprise at least 40 per cent of the electorate. The optional law referendum provisions leave the voters without any means of demanding that a passed bill be submitted to a referendum. If a minority in the Folketing puts forward such a demand, the constitutional provisions clearly favor representative democracy over direct democracy. What is left of direct democracy is in realty only a veto, a right to block bills passed by the Folketing.

History of Referendums

Denmark has experienced more national referendums (17 so far) than most countries [Note that the article was written in 1995. Since then and up to now, April 2002, there have been two more referendums, one on the Amsterdam treaty, 28.5. 1998, 74.8% turnout, 55.1% yes, 44.9% no, and the other one about joining the Euro-zone, 86.6% turnout, 46.8% yes, 53.2% no. Moreover, the number (17) does not include partial referendums - those confined to part of the territory - for example, the vote in the West Indian Islands in 1868 or the vote in Greenland in 1978 about prohibition on liquor rationing. Neither are the votes in Nord Schleswig in 1920, in Iceland in 1944 and on the Faeroe Islands in 1946 - regarding the nationality of these provinces - included].

The frequency of referendums is increasing - 14 of the 17 referendums have been held in the period since 1953. However, it is somewhat misleading to state - as has been done - that they have been just as common as general elections. It is true that in the period from 1953 to 1980 11 referendums and 12 general elections were held, but four of these referendums were held at the same time (in June 1963 on land reforms) and two of the remaining ones were also held at the same time (in May 1953 on a constitutional amendment and a lowering of the voting age). There were, accordingly, seven "referendum occasions" in the 1953-80 period, and one of these (in 1971) was combined with a general election. For the 1953-95 period, 18 general elections should be compared not with 14 referendums but rather with 10 "referendum occasions". In Table 3.1 the referendums to have been held are classified according to the principal types outlined above.

Danish referendums could also be classified according to the issues dealt with rather than the constitutional type. With a few exceptions, they fall into three main categories: the constitution, voting age, and the European Community.

Constitutional Amendments

Through an active effort, even smaller parties and other political groups may be able to jeopardize the passage of an amendment agreed by the larger parties. In 1953, as we have seen, the four old parties and the Single-Tax Party were in favor of the constitutional amendments, the Communists being the only party represented in parliament that opposed the proposed amendments. In addition, there was some opposition in bourgeois circles - especially among the Liberals. These political forces from the extreme left and the right succeeded in mobilizing over a fifth of the votes cast against the proposal. The agreement among the larger parties was reflected in the highest turnout to date in a referendum on constitutional amendment, but it was still about 20 per cent lower than at the previous general election, and the requirement of 45 per cent of the electorate was only just achieved.

Mandatory Referendums on the Voting Age

Bills concerning reduction of the voting age have been a recurrent theme in postwar political debate - five of the 17 referendums concern this question. Apart from the referendum in September 1971, which was held at the same time as a general election, the turnout at voting age referendums has been over 20 per cent below that at the nearest general election. In 1961, in fact, turnout was almost 50 percentage points below the turnout at the general election the previous year. The lower turnout could be explained either as an expression of limited interest in this constitutional issue or as an expression of confidence in and acceptance of parliamentary decisions. Bearing in mind the result in 1969 and the rather modest majorities - usually around 55-45 in favor - the first explanation seems more likely. Referendums may have a conservative effect by delaying achievement of reform demands backed by a majority in parliament. In 1953, the electorate chose the higher of the alternative voting ages presented to them by parliament. In 1961 only about a third of the electors took part in the referendum, but a stronger turnout might well have resulted in an even higher vote against reducing the voting age to 21 years. Finally, in 1969 the voters rejected a bill for a voting age of 18 years - a bill that was, however, approved in 1978.

Mandatory Law Referendums on Delegation of Sovereignty

The third main issue for Danish referendums is the European Community (EC), on which four referendums have taken place. The most dramatic and epoch-making referendum in Danish politics was the vote in 1972 regarding Danish membership of the EC. This referendum was extraordinary for several reasons.

First, the issue was raised in an unusual fashion. As early as May 1971 a Social Democratic leader, Per Hækkerup, suggested that an advisory referendum be held before the final vote in the Folketing. A few days later his party agreed that a decisive referendum on Denmark's membership of the EC should be held according to the rules laid out in Article 20 of the constitution - regardless of whether or not a bill on this topic was passed with a majority of five-sixths in the Folketing. The Radical Liberals concurred with this proposal, which was then hesitantly accepted by the two larger bourgeois parties.

Secondly, this early call for a referendum before it was known whether one would be constitutionally necessary reminds us that referendums may serve other functions than the obvious ones - in particular, they may serve as a device for removing an issue from the political agenda. The Social Democrats - and to some degree the Radical Liberals - were in the early 1970s internally divided on the EC issue, and the proposal to submit this issue to the people whatever the size of the parliamentary majority for EC membership indicates that the Social Democrats wanted to remove the issue from the forthcoming election campaign. By referring the EC issue to a later referendum, the party hoped to prevent a loss of voters to the Socialist People's Party, which strongly emphasized the EC issue and was opposed to Danish membership.

Thirdly, the 1972 referendum was extraordinary due to the campaign and the result of the vote itself. The campaign leading up to the referendum in October 1972 was undoubtedly the most comprehensive and enthusiastic ever to take place in Danish politics. lt was characterized by a confrontation between "the establishment" - the leaders of the old parties and the large organizations - on the one hand and the parties of the left, new organizations and the grassroots on the other. This campaign contributed to internal difficulties in some parties - in particular the Social Democrats and the Radical Liberal party - and it furthered the conception of a cleavage between the politicians and the people as well as a belief in a crisis of democracy. The turnout of 90 per cent - larger than at any general election - demonstrates the high popular interest in the EC issue and the importance attached to the referendum.

Aspects of this referendum contribute to the debate over whether referendums function as a conservative guarantee for minorities. The result demonstrated a significant difference between the members of the Folketing, of whom almost 80 per cent were in favor of Danish membership, and the voters, of whom only 63 per cent were in favor (see Table 3.1). But this was not, obviously, a case of a small majority in the Folketing passing a bill not approved by the voters - opponents of membership were far from constituting a majority of the votes cast. Accordingly, the vote had no conservative effect in the sense of having a retarding consequence for political development. With regard to party politics, the voting might be said to have had a conservative character as support for Danish EC membership increased strongly from the left to the right of the party spectrum.

The second referendum on the EC occurred in February 1986, when the Danes voted on the Single European Act (SEA). The parliamentary position of the bourgeois coalition minority government formed in 1982 was an important part of the political background. On several issues the government was confronted by a so-called "alternative majority" formed by the Socialists, Social Democrats and the Radical Liberals, the center party that tipped the balance between the Socialist opposition and the bourgeois government. One important manifestation of the power of the "alternative majority" was the rejection by the Folketing of the SEA, which had been approved by the government during negotiations with other EC governments.

Instead of calling new elections to the Folketing on the SEA the government preferred to call an advisory referendum. This illustrates another function that referendums may have in a parliamentary system - not only can they be used to remove an issue from the agenda (as in 1972), but they may also serve to resolve a parliamentary deadlock. The outcome of the referendum supported the view of the government, and with a turnout of 75 per cent the result was respected by a majority in the Folketing, which then approved the SEA.

The third Danish referendum on the issue of European integration was on the Maastricht Treaty in June 1992. This was a mandatory referendum, since the necessary five-sixths parliamentary majority for delegation of sovereignty was not obtained in the Folketing. However, the government had clearly indicated in advance that a referendum on Danish participation in the European Union would be held in any case.

Opposition to the Maastricht Treaty came from different social and political groups, especially from the left and right wings of the spectrum - though for quite different reasons. The left (the Socialist People's Party) in particular attacked the military and foreign policy perspectives of the treaty, as well as the prospect of a common currency. It also tended to see the social dimension as too weak. The two opposition parties to the right of center rejected the treaty for other reasons. The Christian People's Party was divided internally on the issue, but the opponents of the treaty mainly attacked the military and foreign policy elements, being less concerned about the common currency and the social dimension. The Progress Party on the extreme right focused on the social dimension, talking about it as a kind of centralized socialism. It also rejected the common currency, while having no particular objections to the military and foreign policy cooperation in the treaty.

Even this combination of reasons for rejecting the Maastricht Treaty, complicated enough as it may seem, did not give the full picture. The three anti-Maastricht parties succeeded in rallying far more backing for their cause than their regular supporters - in the previous election to the Folketing, in December 1990, the Socialist People's Party had received only 8.3 per cent of the votes cast, the Christian People's Party 2.3 per cent and the Progress Party 6.4 per cent, and the bill on Danish ratification of the treaty was passed by an overwhelming majority of 130 of the 155 votes cast in the final vote in the Folketing. In the referendum campaign, the anti-Maastricht parties were strongly bolstered by spontaneously organized popular movements.

The referendum resulted in a narrow No majority (see Table 3.1). This outcome was the strongest expression in modem Danish history of a deep divergence between the political elite, who overwhelmingly backed the treaty, and the voters, who rejected it. The discrepancy between the elite and the voters is comparable to the 1969 referendum on the voting age, but on an issue that is generally acknowledged to be much more important and in which, as shown by the high turnout, the voters take a much stronger interest.

The immediate reaction of the political elite was one of shock and confusion. The government could hardly believe that it had been overruled by the people, and found itself in a difficult situation. How would it be possible to reconcile the resistance among Danish voters with the expectations of other European governments and at the same time ensure full participation for the Danish government in the EU? This dilemma caused by the June referendum was solved by the so-called "National Compromise" among seven of the eight parties in the Folketing, the Progress Party being the only outsider.

In October 1992 the Socialist People's Party invited the Social Democrats and the Radical Liberals to discuss the possibility of a compromise on the Maastricht Treaty. Within a few days the three opposition parties had agreed on a document that the government more or less had to accept. The main content of the "National Compromise" concerned a demand for four exemptions from full Danish participation in the EU [The four exceptions are 1. Denmark does not participate in the so-called defense policy dimension involving membership of the Western European Union and a common defense policy, 2. Denmark does not participate in the single currency and the economic policy obligations linked to the third stage of the EMU, 3. Denmark is not committed in relation to Union Citizenship, and 4. Denmark does not accept transfer of sovereignty in the area of justice and police affairs].

The European Council at its meeting in Edinburgh in December 1992 reached an agreement that to a large extent sanctioned these exemptions. Even though a sufficient five-sixths majority was obtained in the Folketing for the Edinburgh Agreement, a binding referendum was nevertheless arranged. This provision was part of the "National Compromise", which also included a clear statement that all future major changes in Denmark's relationship with the EU must be subjected to referendums in order to secure legitimacy. The fourth Danish EC/EU referendum thus took place in May 1993. Among the parties the Edinburgh Agreement was opposed only by the Progress Party, but several popular movements campaigned against it. The result showed popular approval for the Edinburgh Agreement, but the anti-EC/EU resistance was once again strongly demonstrated, the No vote amounting to 43 per cent of the votes cast.

Optional Rejective Law Referendums

As described above, the possibility of holding an optional law referendum was introduced in the 1953 constitution, but this opportunity has been utilized only once, in June 1963. The Liberal and Conservative members of parliament demanded a referendum on four land reform bills passed by the Social Democrats, the Radical Liberals and the Socialist People's Party. The bills contained a number of principles concerning the balance between private property on the one hand and a higher degree of public regulation on the other. The consistency of the Yes and No votes on all four bills suggests that the voters saw the four proposals as involving one central issue - only the bill on nature conservation dissociated itself a bit from the rest with a slightly lower No vote (see Table 3. 1).

The four bills were rejected by the voters, with both conditions for rejection being met: a clear majority of the votes cast were against the bills, and the No votes comfortably exceeded 30 per cent of the electorate. Before 1953 doubts had been expressed about the applicability of the new referendum provisions, which it was claimed would not be able to operate as intended. Leading lawyers had described the provisions as mere window-dressing. What happened in 1963 showed, however, that the optional referendum could serve as an effective minority guarantee. The turnout was higher than at any previous referendum and only 12-13 per cent below the turnout at the previous and the following general elections. The size of the majority left no doubt about the population's attitude towards the bills. Even sceptics had to admit that this vote was a strong argument for referendums as a part of Danish democracy.

As a result of the 1963 referendums, efforts in progress on physical planning were postponed and there was a certain change in the principles that lay behind the bills. The bourgeois parties had directed their campaign against the increased scope for expropriation in particular. When the Liberal-Conservative-Radical coalition government in 1969-70 proposed and secured the passage of a number of bills in this area, a less centralized planning model was applied and the word "expropriation" was avoided. However, it could be argued that the body of laws in general aimed at being just as strong as if a wide scope for expropriation had been included, indicating that the bourgeois parties when in government in fact adopted the same policies as the Social Democratic government.

Advisory Referendums

Finally, Denmark has held two advisory referendums, and several more have been proposed. The second one was on the SEA in 1986, discussed above, while the first took place as early as 1916. After negotiations with the United States the Radical Liberal government proposed to sell the Danish West Indian Islands to the USA (Today known as the Virgin Islands). The opposition had a majority in the upper house - the Landsting - and it used the occasion to demand that the new 1915 constitution be put into force and to call for new elections to the Folketing. However, after prolonged negotiations the parties agreed to avoid a general election during wartime. Instead a coalition government was formed and the issue of the West Indian Islands was submitted to an advisory referendum. This is another illustration of a referendum resolving a parliamentary deadlock. The Conservatives started an energetic campaign against the sale, but the turnout was low, and a majority of the votes cast were in favor of selling the islands, a step that was then taken

Since 1916 there have been various suggestions for holding advisory referendums one researcher concluded that referendums were "seriously considered or suggested" no fewer than 21 times from the end of the second world war to the beginning of the 1980s. In more recent years there have been further such suggestions. For instance, in 1989 the prime minister, Poul Schlüter, expressed the view that a comprehensive governmental plan on the economy should be decided in a referendum. More often, however, it has been opposition parties that have launched demands for referendums.

Electoral Behavior at Referendums

We shall examine both whether the voters participate in referendums and how they participate, i.e. how they make a choice between alternative policies.

In Denmark turnout has rarely been higher at referendums than at general elections. One case of high turnout (regarding the voting age in 1971) is explained, as we have mentioned, by a general election being held on the same day. The EC referendum in 1972 had the highest turnout ever, and turnout was also high at the referendums on the land reforms, on the Maastricht Treaty and on the Edinburgh Agreement. In all the other cases, especially when the subject has been a constitutional or quasi-constitutional issue such as the voting age, turnout has been much lower than in general elections - voters are evidently more keen to participate in referendums when specific issues involving important choices are presented to them.

Denmark's referendum voting experience does not support the party-driven voter model. In particular, electoral behavior at the two EU referendums in 1992 and 1993 showed clearly that large numbers of voters do not follow the advice given by their political party. Supporters of the Socialist People's Party are a striking example. When the leadership of this party changed from No to the Maastricht Treaty to Yes to that treaty as interpreted in the Edinburgh Agreement, its lead was followed by very few of its usual voters. Four out of every five voters who had supported the party in the 1990 Folketing election did not follow the advice of its leaders, but voted No once again. The Social Democratic leaders were only slightly more successful: the changes represented by the National Compromise and the Edinburgh Agreement convinced enough Social Democratic voters that whereas two out of three voted No in June 1992 only half of them did so in May 1993. In fact, a close relationship has developed between the voters' attitudes towards European integration and their behavior at the referendums.

The Danish No to Maastricht in 1992 is explicable mainly by the importance attached to political freedom and a fear of losing political sovereignty and independence. More than eight out of ten of those who prefer political freedom to economic benefits if forced to make a choice said No, while the pattern was reversed among those preferring economic benefits to political freedom. The general attitude towards European integration among Danish voters is that economic cooperation and integration is accepted, but political cooperation and integration is rejected.

The Impact of Referendums

Referendums may be said to have had only a limited impact on Denmark's political development. The most important consequences have resulted from the rejection of the land reforms in 1963, the approval of Denmark's entry into the EC in 1972, and the rejection of the Maastricht Treaty and the approval of the Edinburgh Agreement in 1992-93.

Most of the 17 referendums have concerned constitutional issues of limited interest to the voters and they have had a rather indirect impact on actual political polices. However, the EC/EU issue has increasingly been subject to referendums with important political implications, and additional referendums on this subject are likely to follow.