where the majority requirement for required referenda is specified, scholars propose a ‘double’ ma-
jority of voters and of states. Non-required referenda are proposed to be triggered either by voters
or by the EP. Hug criticises that the proposals are not discussed in comparing them with the existing
decision-making procedures in order to find out to what extent the introduction of referenda and ini-
tiatives actually changes policy outcomes in the EU. But he abstains from proposing any blueprint
direct democratic reform of EU decision-making. In Feld, Kirchgassner and Weck-Hannemann
(2002), it is suggested to consider the introduction of ‘budget referenda’ in EU decision-making. On
the one hand, the analysis follows the proposal of Hug by looking more closely on the impact of re-
quired referenda on budgetary outcomes in the EU, given its current budgetary process and the po-
tential process after Eastern enlargement. On the other hand, the authors are not too realistic about
the institutional provisions shaping the budget referendum. They suggest a general required referen-
dum on the EU budget draft at the end of the current budgetary process with interactions of the
Commission, the Council and the EP, totally abstracting from the transaction costs that this proposal
involves. Naturally, a required referendum on the whole budget draft would unduly increase the time
needed to pass a budget.
In this paper, our aim is to analyse to what extent direct democracy helps to resolve the democratic
deficit at the EU level and to develop a European demos. In addition, we propose the introduction of
referenda and initiatives in EU decision-making by going a step further towards realism. We consider
direct democratic decision-making basically in all EU decisions, but develop proposals for its intro-
duction only when it is feasible without prohibitively increasing decision-making costs. In Section 2,
we clarify that each constitution is a contract joining the citizens of a state and requires as such the
explicit agreement of (a majority of) citizens. In addition, we generally discuss the rationale of refer-
enda and initiatives as instruments of control in a representative democracy and summarise empirical
results on the impact of direct democracy on policy outcomes in Switzerland and the U.S. These ar-
guments imply two different proposals for the EU in Section 3: First, the future European Constitu-
tion should also be approved by the European citizens once the Convention has proposed and the
Intergovernmental Conference has agreed on the provisions of the Constitution. A basic consent of
the European people is required to found a new federation.3 Second, future changes of the Constitu-
tion have to be decided by citizens in European referenda as well. In these two cases, we propose a
mandatory (required and binding) referendum on total and partial revisions of the European
Constitution. Third, a pre-specified number of citizens from a pre-specified number of countries
should have the right to initiate constitutional changes at the EU level. In this case, we thus propose a
popular initiative on partial revisions of the Constitution.
Given the current state of affairs, most of European politics would be covered by the proposed con-
stitutional referendum and initiative. A look at the ‘skeleton’ for a EU constitution proposed by the
Convention reveals however that only the first part contains provisions that are considered constitu-
tional in the traditional sense. The second part of the draft treaty is supposed to contain statutes on
the accomplishment of EU policy measures while the third part includes provisions for a continuity
from the current Treaties to the new Constitution as well as on the adoption and revision of the Con-
stitution. Requiring a referendum on the second part would restrict EU politics too strongly. We
3. See also: Let the people vote, in: THE ECONOMIST 367 (8325), May 24, 2003, pp. 10.