Which countries only have direct democracy?

Direct democracy: engine or brake?

How referendums delay European unification

by Uwe Wagschal

Two years ago the European constitution failed in the French and Dutch referendums. Is direct democracy the engine or the brake on European integration? Which factors influence national referendums on European policy content? A research project by the Institute for Political Science at Heidelberg University has received illuminating answers to these questions.

European integration steps have already been rejected by EU member states in referendums in the past. Examples of this are the Danish "No" to the Maastricht Treaty (June 2, 1992), Denmark (September 28, 2000) and Sweden (September 14, 2003) opting out of the euro area and the Irish rejection of the Nice Treaty (8 September 2003) June 2001), which at the time also threatened the accession of the ten new member states in 2004. In substance, however, these decisions were less problematic than the failure of the Treaty establishing a Constitution for Europe (TCE) or - as in the case of Ireland - could at least be turned in retrospect through a further vote.

Whether direct democracy is to be interpreted as a powerful veto institution with a great "political braking effect" can only be determined by comparing it with the procedures of representative democracy. The ratification of European treaties cannot only be "jeopardized" by referendums. For example, a high majority requirement in parliament (e.g. a two-thirds majority) may be necessary or the opposition may control a second parliamentary chamber, which could also prevent ratification. In general, the ratification process of the draft constitution in the 27 member states of the European Union envisaged three different paths: 17 countries wanted to adopt the constitution only through parliamentary channels, while five countries provided for it to be passed by parliament as well as by the people. A passing only by the people was also planned in five countries. Eight countries have currently postponed the ratification process or have not yet completed it.

When will there be a referendum? In the case of mandatory referendums, a government has no way of influencing the political agenda, except on the date of the vote. On the other hand, in the case of optional and consultative referendums, it always has leeway to schedule a vote or to prevent it. Governments can influence the ratification process through their definition competence: For example, the Lithuanian government prevailed with its view to classify the TCE as an international treaty and not as a constitutional amendment. For this reason, "only" the Lithuanian parliament had to approve, otherwise a mandatory referendum would have been necessary. Similar considerations shaped the behavior of Italy, where a referendum was also avoided.

The table above shows the ratification procedure and the status of the ratification process as of January 1, 2007. As you can see, at this point in time 15 of the 17 states in which only one parliamentary approval was planned had already approved the EU treaty. In Germany, only the signature of the Federal President of the ratification law passed in May 2005 by the Bundestag and Bundesrat with a two-thirds majority was still pending, but this was delayed by a pending lawsuit before the Federal Constitutional Court. Sweden, on the other hand, postponed ratification indefinitely. If one looks at the countries that provided for a mixed ratification by parliament or referendum, a more differentiated picture emerges: two countries had passed the constitution both by referendum and by parliament (Spain and Luxembourg). The outcome of the procedure in the Netherlands was negative, while Ireland and the United Kingdom had not yet scheduled voting dates. None of the countries that envisaged ratification only by referendum have not yet completed the process: In addition to the failed vote in France, as of January 1, 2007, the votes in the Czech Republic, Denmark, Poland and Portugal were still pending.

In the parliamentary procedure, support was never denied to ratification. This also applies in the case of high approval hurdles, opposing majorities in the second parliamentary chamber or if there is a high incentive for the opposition to inflict a severe defeat on the government. Obviously, the referendums created stronger time delay and blockade effects in the European integration process than the strict regulations that most countries apply for ratification in the parliamentary procedure.

Economic and political framework

If you expand the focus to national referendums with a European reference, you can identify a total of 52 referendums between 1955 and 2006. In the course of a research project on the effects of direct democracy carried out at the Heidelberg Institute for Political Science, in addition to the voting results, the economic and political framework conditions at the time of the voting were surveyed and the public debates were analyzed in more detail.

The first Europe-relevant referendum after the Second World War took place in Saarland in October 1955 and its subject matter was Europeanization. There are also three non-European referendums among the 52 referendums: one on Greenland and two in Azerbaijan, as well as a further, sub-national vote on the Finnish Ă…land Islands (1994). Summing up the results, a total of 38 votes were accepted and 14 were rejected. With regard to the national referendums in the most important European countries - the 27 EU countries and the four EFTA countries - there are three countries with Norway (1972 and 1994), the Netherlands (2005) and Cyprus (2004) in which only rejected decisions were observed. With Sweden, Ireland, Denmark, France and Switzerland, there are at least five countries that have both negative and positive referendums. The group of six EU or ETFA countries in which there have not yet been any European votes (Belgium, Bulgaria, Germany, Greece, Iceland and Portugal) represent only a minority. After all, 17 countries have only passed positive referendums.

Which topics were up for debate in the decisions examined? The votes can basically be divided into four blocks, which, however, have to be drawn up relatively broadly due to the different national legal situations. Most of the votes (n = 26) were on joining or leaving (Great Britain 1975 and Greenland 1982) European institutions. The second largest group of European-relevant votes deals with European treaties (n = 19) such as the Maastricht, Amsterdam and Nice treaties or with the VVE. National constitutional reforms (n = 4), which sometimes preceded accession to European institutions, or territorial agreements (n = 3) only make up a minority. If one looks at the development of direct democratic votes over time, an increase in this set of instruments can be observed: half of all votes (26 of 52) took place after the year 2000.

How can the different proportions of approval - measured by the proportion of yes votes in percent - be explained in the votes? Nine different hypotheses and theories were used to explain and checked for their validity with the help of statistical analyzes.
  1. First of all, it is to be expected that countries that can expect relatively high losses in prosperity are more likely to reject European provisions. Conversely, the proportion of votes in favor is likely to be higher in those countries that are particularly poor and economically expect a lot from accession. This rent-seeking thesis is clearly confirmed in the figure on page 41, which shows the relationship between the proportions of votes in favor and the "prosperity gap" of a country to the OECD average. The statistical outlier Luxembourg is noteworthy, which shows "too high" approval measured by its level of prosperity. However, this also fits very well with the rent-seeking thesis, as Luxembourg benefits greatly from the EU and its institutions.
  2. The identity thesis states that the likelihood of approval increases the higher the European identity or approval of Europe as a whole. This identity thesis can be checked against the approval of Europe measured in the Eurobarometer surveys. The statistical relationship between the proportions of yes votes and the questioned agreement is medium (r = 0.390; n = 38), but it is consistently significant in the multivariate regression analyzes.
  3. The first variant of the participation thesis states that approval falls with an increasing number of voting options in a country (rules-in-form), because the citizens are "oversaturated" with many votes. This effect can be observed above all in the "motherland of direct democracy", Switzerland. In the second variant, the thesis is supplemented by the factual participation in the vote (rules-in-use). However, both formulations are not confirmed.
  4. The second-order thesis is more explanatory. It is based on the idea that referendums are not only used for their actual purpose - the vote on the matter - but also as a "penal instrument". Such a strategic deployment can be seen well in European and state elections, for example. In the European elections of June 2004, for example, the national governing parties of the then 25 EU countries had to experience an average loss of votes of 11.5 percentage points compared to the previous national parliamentary election. This thesis is checked on the basis of the time interval between the vote and the previous national parliamentary election. The indicator turned out to be significant in the statistical regression analyzes.
  5. The discourse thesis was also confirmed in the analyzes. This thesis postulates that the approval of the referendum decreases the more the political actors argue. One factor for successful referendums is therefore the persuasion strategies and the ability of the elites to reach consensus. The conflictual nature of the domestic political discourse was rated with 3 = "high or disputed", 2 = "medium or normal" and 1 = "low or unanimous". Evaluations of the voting campaign in the yearbook of European integration, in the archive of the present as well as in contemporary or current journalistic reporting in Germany and partly in the countries under investigation served as the basis. If the countries with a low or medium level of conflict are summarized, the average agreement is over 66 percent, while in countries with a high conflict level the proportion of votes in favor was only around 50 percent.
  6. An important object of investigation in political science is how economic outcomes and social problems influence the wishes of voters. Democratic systems legitimize themselves to a large extent through their political output. As a result, the more citizens benefit from a system, for example in the form of low unemployment or a low inflation rate, the more legitimacy is given to a system. Conversely, it can be formulated as a problem pressure thesis that approval declines with growing socio-economic pressure. In the empirical analyzes, however, exactly the opposite occurs. This in turn can be seen as support for the rent-seeking thesis: Obviously, the EU is seen here as an opportunity and a problem-solver.
  7. In terms of the autonomy thesis, it would be expected that referendums with a European reference would tend to be rejected in federal and ethnically fragmented systems. Striving for autonomy and self-reliance and the desire to preserve cultural identity could in fact also go hand in hand with a vague discomfort with Europe. The thesis was measured on the basis of a country's degree of federalism, but could not be confirmed.
  8. The party difference thesis could provide a further explanatory factor. It was checked whether the approval rate is higher among left or right-wing governments, with left-wing parties assuming stronger support for Europe. It turns out, however, that the party difference does not provide any explanatory contribution in any of the test variants.
  9. Finally, the influence of institutions was checked. The institutionalist thesis assumes that the proportion of votes in favor increases both with the number of parties in the government and with their stability. In addition, tests were carried out to determine whether membership in the EU itself or its specific duration had an effect. However, all assumptions prove to be inexplicable.
The central theses are summarized again in the table on page 40 and evaluated on the basis of their empirical validity. The central key factors for explaining the voting results are primarily of an economic nature: voters in rich countries expect a loss of prosperity, while voters in poor countries hope for prosperity gains. The economic situation of a country thus also shapes the individual voting behavior of its citizens. A similar interpretation also applies to the problem pressure thesis. The worse a country is doing economically, the greater the approval of its citizens for Europe. The attitude of the citizens is also an important criterion: the more positively the citizens are towards European unification, the more likely they are to vote "yes". Another decisive factor is how vigorously the domestic political discourse is conducted: the more the political parties argue, the lower the citizens' approval of the referendum. But the second-order thesis also has a certain influence - voters apparently tend to punish their governments on European policy issues if the last election at national level was a long time ago.

When analyzing the ratification of the European Constitution, it must first be noted that it is actually not enough to focus the question on "direct democratic votes versus representative democratic adoption". There is also considerable potential for blockade in parliamentary institutions and procedures, for example when second chambers or high quantitative majority requirements come into play. But even if you take these factors into account, it becomes clear that the ratification of the Treaty establishing a Constitution for Europe is progressing much more slowly in those countries that allow the people to decide, and that this procedure ultimately even led to the rejection of the draft constitution in two cases has led. All in all, the answer to the question of whether direct democracy is the engine or the brake on European integration is clear: referendums have significantly slowed the process of European unification.

Prof. Dr. Uwe Wagschal has been Professor of Comparative Government Studies at the Institute for Political Science since October 2005, which he currently heads as Managing Director. Before that he was Professor of Empirical Methods in Political Science and Policy Analysis at the Ludwig Maximilians University in Munich.
Contact: [email protected]
Telephone: 0 62 21/54 28 68