The Rise of ‘Politically Obligatory' Referendums: The 2005 French Referendum in Comparative Perspective

Abstract

This article compares the use of the referendum in France by the President of the Republic with its use at the initiative of governments in other democracies. It strengthens the current view of a specificity of French referendums, consisting in their power-reinforcing use by the president, and relates it to the personal initiative and the French institutional culture. But French referendums have had other functions as well, like ensuring the adoption of legislation or increasing its legitimacy. Moreover, the use of the referendum for a variety of political ends is a dominant feature of government-initiated referendums in all countries. A straightforward relation between the type of initiative and the type of use of the referendum cannot however be drawn, just as between the initiative, or the use, and the democratic quality of the referendum. The final section focuses on the ‘politically obligatory’ referendum: from being a political resource in the hands of governments, government-initiated referendums are increasingly forced upon them. Two different reasons – appropriateness and pressures – have been found to be conducive to such a political obligation. The decision by the French president to hold a referendum on the EC Constitutional Treaty, which is closely analysed, provides a clear example of a politically obligatory referendum, in contrast to the traditional entrepreneurial use of the referendum by the president. Some decisions by other governments to submit the treaty to the people can also be interpreted in this way.

Notes

1. According to several authors, a general theory of why referendums occur is lacking and almost impossible as each referendum would be the unique product of specific historical conditions (Lijphart 1984: 204–6; Gallagher 1996: 235; Von Holsteyn 2006).

2. We refer here to the comprehensive approach developed by Max Weber (1971[1913). On the relation between intentions and explanation in social sciences, see also Brown (1998).

3. For a typology linking the type of initiative to the extent of popular control on legislation, see Uleri (1985: 246).

4. ‘A continuum running from “controlled” through “uncontrolled” could therefore alternatively be labelled in terms of desired or intended effect’ (Smith 1976: 6).

5. Government-initiated referendums are to be distinguished from constitutionally obligatory referendums, on the one hand, and referendums initiated by minorities, on the other. Minorities initiating referendums may be popular (a certain number of citizens) or institutional (e.g. one-third of the members of Parliament in Denmark, or five regions in Italy).

6. His classification of referendums is indeed the conclusion of an empirical study dealing with a small number of cases including all kinds of formal initiatives.

7. Should the second hypothesis be confirmed, we must admit, however, that the current view according to which government-initiated referendums would be more liable to political use than other referendums would still need to be demonstrated.

8. Again, Smith could be mentioned here as he argues that most government-controlled referendums are ‘pro-hegemonic’ for the government and the regime, while most non-government-controlled referendums are ‘anti-hegemonic’ (Smith 1976: 6, 9). However, it should be noted that Smith does not explicitly relate hegemonic effects to a lower democratic quality of government-controlled referendums – this would suggest that not only the initiative but also the whole referendum process is government-controlled. Rather, his explanation is that governments organise referendums only when they are almost certain to win them.

9. Smaller countries like Andorra, San Marino or Liechtenstein have not been included, although some have had an intensive practice of referendums.

10. This was done through case studies. We are aware of the difficulty of establishing for certain what an actor's goal was, but nonetheless we managed to collect reasonable evidence, relying on a sufficient number of converging indications, for most referendums. Of course, it is not rare that premiers have had several goals, as reported in our examples. But it was generally possible to bring out a primary goal of the referendum. With the exception of France, decisions to hold referendums have been collective decisions, given the necessity of a law enabling the referendum to be passed in parliament. The focus on premiers might thus be regarded as a limit to the capacity of this approach to constitute a first step in the explanation of why referendums occur, especially in the case of coalition governments. It has appeared however that the prime minister's goal was often decisive, the referendum being imposed by him on the parties in the coalition. It should be added that many government-initiated referendums have been decided by prime ministers of single-party governments. This should not come as a surprise as prime ministers do enjoy much greater room for manoeuvre in this case.

11. For more details regarding the examples mentioned in this section and the following, see Morel (1996a and 2001a).

12. A constitutional revision according to which any further enlargement of the EC will require a referendum was voted in 2005 at the initiative of the President of the Republic.

13. According to Wilsford (1994: 251), in situations of path-dependency change can occur only thanks to special conjunctures, or ‘occasional windows of exceptional opportunity’.

14. By removing the issue from the electoral platform, Gonzalez could hope both to attract pro-NATO electors and to escape the future responsibility for a decision to withdraw from NATO.

15. Longhurst's article is particularly worth mentioning here as it is a path-dependency analysis of a security policy, namely the permanency of conscription in Germany. In this specific case, political entrepreneurship and cultural change are regarded as two necessary conditions for breaking path-dependency. We could indeed argue that another – long-term –condition for change in Spain on the NATO issue would have been cultural change and generational renewal on the Left. Gonzalez's entrepreneurship and his use of the referendum have played the role of an accelerator.

16. The word ‘plebiscite’ is used in the literature with a plurality of meanings. However, its most current present acceptation refers to the personal use of the referendum.

17. Mitterrand and Chirac even declared that they would not resign in case of a victory of the No vote.

18. Since 1994 in Iceland the President of the Republic can initiate a referendum on a law that he refuses to promulgate. In Australia, the General Governator (who is the representative of the monarch) can initiate a referendum on a project of constitutional reform about which the two Chambers disagree. These provisions were never used however in any of these two countries.

19. The prime minister either proposes the referendum to the President of the Republic (article 11) or countersigns the presidential decree enabling the referendum (article 89).

20. The 1958 referendum on the Constitution of the Fifth Republic resulted from a common decision of the executive and the legislative powers, as it was proposed by De Gaulle and a referendum law was then voted by the parliament. All referendums held under the new Constitution were initiated by the executive alone.

21. Only Italy (1946) and Greece (1974) have had referendums initiated by governments on their own. In both cases, however, the referendum took place in a context of regime transition. These referendums can actually be compared to those held in France in 1945 at the initiative of De Gaulle, who was the head of the gouvernement provisoire.

22. The discussion of this point would take us beyond the scope of this study. We will limit ourselves to note that governments may have superior means, especially campaign means, at referendums not initiated by them as well. The independent variable, then, would be less the initiator than the rules surrounding the practice of referendums (especially campaign regulations) and, ultimately, the democratic quality of the regime.

23. Our aim here is not to provide a definitive answer to this question – this would require an extensive empirical investigation into the democratic quality of government-initiated referendums. It is only to set the problem and to indicate a few reasons why the political use of referendums by governments should not necessarily be seen as conducive to bad democratic practice, and cannot be regarded, in any case, as the only factor leading to this result – not even the most important.

24. This is the name given to potential candidates to the Presidency of the Republic in France.

25. All the more since the referendum was preceded by internal referendums within the main parties.

26. On voting behaviour at referendums, see the special issue of European Journal of Political Research, 41:6 (October 2002).

27. In a legitimation referendum, the government might also feel constrained to hold the referendum in so far as it might appear as the only way to secure sufficient legitimacy for a decision. Similarly, a referendum might be seen by a government as the only way to avoid a party split (mediation function), the discussion of an issue during a campaign (agenda function), to pass a law (legislative function), or even to rescue its power (power-reinforcing function). In all these cases, one could probably also speak of ‘de facto obligatory referendums’. However the politically obligatory referendum that we are discussing here remains something different as the constraint experienced by governments consists in a pressure to hold a referendum, with the consequence that the aim of the referendum is also specific (avoiding sanctions for having resisted this pressure). It should be added that de facto obligatory referendums aimed at solving the problems just mentioned have rarely been found in the cases studied, except for legislative referendums, which were often the last solution adopted by governments after having contemplated various alternatives to get a law passed.

28. The German sociologist mentioned himself the limits of this interpretation: in particular he insisted on the role of emotions in decisions taken by individuals (Weber 1971 [1913]: 329).

29. ‘By “rules” we mean the routines, procedures, conventions, roles, strategies, organizational forms, and technologies around which political activity is constructed. We also mean the beliefs, paradigms, codes, cultures, and knowledge that surround, support, elaborate, and contradict those roles and routines’ (March and Olsen 1989: 22).

30. The 1905 referendums on the dissolution of the union with Sweden and on the choice between Monarchy and Republic.

31. The referendum is constitutionally obligatory if the treaty has been adopted by less than five-sixths of the parliamentarians. In 1993 a law was voted according to which a binding referendum is allowed even if this number is reached.

32. By contrast, the two other prominent candidates at the presidential election, Ségolène Royal and François Bayrou, both committed to a new referendum on the EC Constitutional Treaty.

33. This has reflected Blair's ability to keep at least the control of the timing of the referendum, which is another key aspect of the control of the referendum. Chirac was much less successful in this, as he could not postpone indefinitely the date of the referendum.