ARENA
Working Papers |
|
Number | Country | Pop. size | (1,2,5) | (1,2,5) | Pres. w. | ||
4 | Germany | 81.2 | 25 | 12 | 10 | ||
UK | 58,4 | 25 | 12 | 10 | |||
France | 58,0 | 25 | 12 | 10 | |||
Italy | 57,2 | 25 | 12 | 10 | |||
1 | Spain | 39,2 | 18 | 9 | 8 | ||
4 | Netherl. | 15,4 | 10 | 5 | 5 | ||
Greece | 10,4 | 10 | 5 | 5 | |||
Belgium | 10,1 | 10 | 5 | 5 | |||
Portugal | 9,9 | 10 | 5 | 5 | |||
2 | Sweden | 8,8 | 8 | 4 | 4 | ||
Austria | 8,0 | 8 | 4 | 4 | |||
3 | Denmark | 5,2 | 6 | 3 | 3 | ||
Finland | 5,1 | 6 | 3 | 3 | |||
Ireland | 3,6 | 6 | 3 | 3 | |||
1 | Luxemb. | 0,4 | 3 | 2 | 2 | ||
EU-15 | 370,9 | 195 | 96 | 87 | |||
Quota | 139 | 69 | 62 | ||||
If, instead, we choose the voting weight distribution (1,3,5) as our EC-6 point of departure, the medium-sized countries will gain. The results are presented in Table 2.
Table 2. Voting weight distributions for EU-15 based on EC-6 (1,3,5)
5* 2,5*
Number | Country | Pop. size | (1,3,5) | (1,3,5) | Pres. w. | ||
4 | Germany | 81.2 | 25 | 12 | 10 | ||
UK | 58,4 | 25 | 12 | 10 | |||
France | 58,0 | 25 | 12 | 10 | |||
Italy | 57,2 | 25 | 12 | 10 | |||
1 | Spain | 39,2 | 20 | 10 | 8 | ||
4 | Netherl. | 15,4 | 15 | 7 | 5 | ||
Greece | 10,4 | 15 | 7 | 5 | |||
Belgium | 10,1 | 15 | 7 | 5 | |||
Portugal | 9,9 | 15 | 7 | 5 | |||
2 | Sweden | 8,8 | 12 | 6 | 4 | ||
Austria | 8,0 | 12 | 6 | 4 | |||
3 | Denmark | 5,2 | 7 | 4 | 3 | ||
Finland | 5,1 | 7 | 4 | 3 | |||
Ireland | 3,6 | 7 | 4 | 3 | |||
1 | Luxemb. | 0,4 | 3 | 2 | 2 | ||
EU-15 | 370,9 | 228 | 112 | 87 | |||
Quota | 167 | 80 | 62 | ||||
(4) The relevance of the above analysis
In the European Community great emphasis is put on l'acquis communautaire. Although this principle primarily applies to other aspects of the organisations, institutional continuity has a high value. In so far as there is a need for increasing the voting weights of the large states, it may be of some interest that the modification suggested on the basis of EC-6 (1,2,5) reflects the thinking behind the set-up of the Europe of Six. It is quite interesting that the Netherlands, in chairing the IGC, proposed a revision very close to my extrapolation of the EC-6 (1,2,5). There are only two differences: The Netherlands proposed voting weight 20 instead of 18 for Spain, and 12 instead of 10 for the Netherlands themselves (the population of the Netherlands is 50 % above that of Belgium). This proposal was not based on considerations related to EC-6 [2]. That means that the voting weight distribution proposed has different qualities that can make it a `focal point' (cf. Schelling 1960) in future negotiations.
It should be noted that the Netherlands' complete proposal included two main alternatives: first, a double majority procedure (in two variants) and, second, a revision of the voting weight distribution. The proposal as a whole is presented in Table 3.
Table 3. The Netherlands' proposal May 1997 under the IGC
Alternative 1:
Double majority:
71,1 % of present votes/weights (viz. 62 out of 87 votes)
60 % of population
Double majority:
Majority of member states
60 % of population
Alternative 2: Revision of weights with majority requirement as now, viz., about 71,1 %
Number | Country | Pop. size | Weight 1 | Weight 2 | Pres. w. | ||
4 | Germany | 81.2 | 25 | 12 | 10 | ||
UK | 58,4 | 25 | 12 | 10 | |||
France | 58,0 | 25 | 12 | 10 | |||
Italy | 57,2 | 25 | 12 | 10 | |||
1 | Spain | 39,2 | 20 | 9 | 8 | ||
1 | Netherl. | 15,4 | 12 | 6 | 5 | ||
3 | Greece | 10,4 | 10 | 5 | 5 | ||
Belgium | 10,1 | 10 | 5 | 5 | |||
Portugal | 9,9 | 10 | 5 | 5 | |||
2 | Sweden | 8,8 | 8 | 4 | 4 | ||
Austria | 8,0 | 8 | 4 | 4 | |||
3 | Denmark | 5,2 | 6 | 3 | 3 | ||
Finland | 5,1 | 6 | 3 | 3 | |||
Ireland | 3,6 | 6 | 3 | 3 | |||
1 | Luxemb. | 0,4 | 3 | 2 | 2 | ||
EU-15 | 370,9 | 199 | 97 | 87 | |||
Quota | 142 | 69 | 62 | ||||
A pertinent question is how much difference it makes whether the voting weight distribution in EU-15 (and in possible enlargements) is the present one, or the ones presented in Table 1 and Table 3 (cf. EC-6 (1,2,5)), or, finally, the one presented in Table 2, based on EC-6 (1,3,5).
Let us note at once:
The Banzhaf power index of a large member state under the present voting weight distribution is 11.16. Under the voting weight distribution we obtain by multiplying the EC-6 (1,2,5) values by 2.5, it is 12.13. Under the voting weight distribution we obtain by multiplying the EC-6 (1,3,5) values by the same number, it is 10.39. (The Banzhaf indices for large states which we obtain by multiplying by 5 instead of by 2.5 are somewhat higher because here there will be no rounding off.)
The Banzhaf power index tells us what the probability is that the country in question will be a decisive member of a winning coalition if all coalitions are equally probable. As was shown in ARENA Working Paper 10/98 (Midgaard 1998), it is equivalent to an index of autonomy, defined as the difference between the likelihood of ending up in a winning coalition and the likelihood of ending up in a losing coalition. [3] The results presented in the preceding paragraph can therefore be given the following interpretation:
- the difference between the number of winning coalitions and the number of losing coalitions that a large state can end up in, will be higher under the scheme of voting weights based on EC-6 (1,2,5), or the one proposed by the Dutch presidency;
- the difference will be lower under the scheme based on EC-6 (1,3,5).
It might be of interest to find out which are the extra winning coalitions, and which are the losing coalitions that are eliminated, under the former scheme. That, however, is beyond the frame of this paper.
(5) Some arguments submitted in favour of increasing the
voting weight of large states
Let us recall from the Introduction:
Leading German politicians (the CDU/CSU parliamentary group), have declared: With regard to the Council, democratisation means striking a better balance between the basic equality of all member states, on the one hand, and the ratio of population size to number of votes in the Council, on the other.
Leading French politicians have asserted that the weighting of votes should be reviewed in order to take into account demographic, economic and financial realities.
What is particularly noteworthy here is that the French politicians do not limit themselves to the demographic realities. That is in fact what the German politicians in question do. The German statement is therefore unproblematic from a democratic point of view. The view expressed by the French politicians in question, however, is not.
Let us then add a view expressed by a leading British politician.
In a Pall Mall Institute publication (1996), the then Labour Shadow Secretary of State for Scotland, George Robertson, later British Minister of Defence, and now the Secretary-General of NATO, makes the following statement:
Enlargement of the EU will bring in more small countries. This will tilt the balance in their favour. But the driving force behind the European Union remains in the large countries. Any new arrangements, however, will have to maintain the predominance of the large member states, but also the built-in safeguard to defend the interests of the small member states. We must recreate the original balance, just with different arithmetic.
This statement, too, is problematic from a democratic point of view.
The three statements constitute an interesting triple. In order to gain perspective on them, let us go back to Aristotle's discussion of different conceptions of fair or just constitutions.
Since Aristotle political equality, i.e. the principle of equal political rights and duties, including the principle of one man, one vote, one value, is one central criterion, or definition, of democracy.
Those who favoured democracy in Greece, in particular the free-born poor, asserted, according to Aristotle, that reason, which is characteristic of all free-born, is the only attribute which is significant when it comes to assigning political rights and duties. This principle competed with two others. The well-to-do tended to assert that those who make significant financial contributions to the state should have more to say than ordinary citizens; and Aristotle maintained that those contribute much wisdom to the polis should have even more to say (Aristotle 1946: III, 9, 1280a). We may altogether talk about three principles of constitutional justice: (1) the principle of constitutional equality, (2) the finance principle of constitutional proportionality, and (3) the competence principle of constitutional proportionality.
All three principles, and principles akin to them, still play a significant role in establishing rights and duties in different kinds of organisations. More specifically, this is true of organisations established by democratic states, or within them. When it comes to the basic institutions of society as a whole, however, the principle of constitutional equality prevails in these states, the central justification being respect for the dignity and worth of the human person.
We need not go very far back to find examples of arguments close to the finance principle of proportionality. Leif Lewin, in his book Ideology and Strategy: A Century of Swedish Politics (1988), gives interesting examples in his account of the debate on suffrage in Sweden, in the second half of the 19th century. For instance, the leading conservative bishop G. Billing maintained that it was only fair that those who pay much more in taxes should have more of a say than those who pay less in taxes or nothing at all. (Lewin 1988, p. 58). It is noteworthy that those who opposed equal voting rights also tended to mean that the well-to-do were politically more qualified and therefore should have more of a say in politics than other citizens.
Now, back to the statements and formulations quoted by representatives of Germany, France and Great Britain, respectively.
It is not unreasonable to suggest that the statement submitted by German
politicians reflects the principle of constitutional equality, i.e. the principle of democracy. Similarly it is not unreasonable to suggest that the view expressed by French politicians reflects the finance principle of constitutional proportionality, i.e. the oligarchic principle. Finally, it is not unreasonable to suggest that the statement submitted by the British politician, Mr. George Robertson, reflects the competence principle of constitutional proportionality, i.e. the aristocratic principle.
Caution is necessary, however. We have to consider the possibility that for each of the statements quoted there is a reasonable interpretation that makes it consistent with the principle of democracy. Not least is there a need to consider whether the French politicians' emphasis on the relevance of economic and financial realities may reflect, or be consistent with, the principle of constitutional equality. To get a better background for our analysis we shall first, primarily on the basis of Stephen Zamora's article Voting in International Organizations (Zamora 1980), have a look at how economic and financial realities have served as a basis for assigning voting weights to actors in international organisations.
(6) Economic and financial realities as a basis for
assigning voting weights in international organisations
The tendency to substitute majority decisions, more specifically weighted voting, for the traditional requirement of unanimity in international co-operation goes back to the second half the 19th century. Stephen Zamora gives the following account:
With the establishment of the International Telegraphic Union (ITU) in 1865, liberal
voting procedures, including the ITU, the Universal postal Union (UPU), the
International Wine Office, the International Office of Chemistry, and the International
Institute of Agriculture, had deviated from the strict one-nation, one-vote rule.
(Zamora 1980, p.575)
He suggests several reasons why the trend to majority began with administrative and technical unions. This is a central one:
...these early organizations were limited in authority and scope to narrow, technical
matters that did not intrude into questions of high national policy. Thus, states felt less
need to protect their interests through a veto power. (Zamora 1980, p. 575)
The number of international organisations that use weighted voting has increased considerably during the 20th century. Central examples are the Bretton Woods institutions, i.e. the World Bank and the International Monetary Fund, and the international commodity agreements, such as the International Coffee Agreement. Far from all international organisations, however, have adopted weighted voting. Zamora submits some interesting reflections on the conditions for weighted voting to be considered a proper decision-making procedure, indicating at the same time mutually conflicting intuitions as to what may constitute a fair distribution of voting weights:
In general, weighted voting can be a workable procedure for organizations
performing specific, well-defined functions that are accepted by their members and
that require their continued support (financial, manpower, etc.) Weighted voting
facilitates the functioning of such organizations because the interests and
responsibilities of the members are not shared equally; therefore, members with
greater obligations have greater opportunities to protect their interests.
This analysis, closely adhered to by proponents of weighted voting, ignores the
fact that in many organizations the relative interest of each member may not be
reflected in that state's financial contribution. For instance, a developing country
member of one of the regional development banks may contribute far less to the bank's capital than the United States, yet the developing country's contribution may be larger than that of the United States when measured in relation to gross national product or per capita income. This consideration, coupled with the importance of the development bank as a lender to the developing country, has far greater interest in the organization than does the United States. The problem, of course, is that the bank
cannot survive without an influx of capital, which usually involves an exchange of votes for dollars (or franks, or marks). (Zamora 1980, pp. 591-92)
Zamora of course also mentions the European Coal and Steel Community and the European Economic Community as examples of international organisations that have adopted weighted voting. He emphasises the lack of explicit and clear criteria for assigning voting weights to the various members of the Council of Ministers.
Unlike the Bretton Woods Agreement, the Treaty of Rome does not indicate the criteria by which the weighted votes are assigned to members. Rather, different numbers are simply assigned to different countries. While the allocation seems to reflect the relative size of populations, it is clear that other factors enter into the determination, since West Germany and France both receive the maximum number of ten votes, but the population of the former is about 15 percent larger than that of the latter. Nor does relative economic power explain the allocation: West Germany and Italy certainly are not equal in this category, the gross national product of the former being two and one-half times of the latter. Rather, the allocation of votes appears to result from a combination of population, economic power, and political reality.
(Zamora 1980, pp. 582-83)
The view that the allocation of votes results partly from population and political reality is hardly surprising. In our context, however, it is interesting to note that in Zamora's view economic power also has played a role. It is clear that the size of the relevant part of the economy plays a role in the decision-rules of the ECSC, although not necessarily for the vote allocation, and it is unproblematic since the ECSC covers two rather narrow, although important sectors of the economy. It is more problematic to make economic power a basis for assigning voting weights in a broad organisation like the EEC, or today's European Union. However, the view that economic power has played such a role also appears in Geoffrey Garrett's article International Cooperation and Institutional Choice: the European Community's Internal Market (Garrett 1992). Dealing with the bargaining over the SEA he says, Thus, the French and German governments preferred a voting system in which super majorities were required and in which votes were weighted in accordance with economic power (Garrett 1992, p. 546).
The way in which French politicians have recently emphasised that voting weights must reflect economic and financial realities among other things, indicates that Zamora's and Garrett's interpretation is not mistaken. Or, to view it from the opposite angle, the interpretation submitted by our two prominent scholars gives us an extra reason for dealing with the French utterances. There is reason to think that the French politicians express considerations related to more than the inclusion of less affluent countries as new members of the Union.
(7) Some arguments pro and con different schemes of
voting weight distributions
Our point of departure in this paper was the following questions: How, and to what extent, should the distribution of seats and votes reflect the equality of the member states, and how, and to what extent, should it reflect the principle of political equality across nations? Moreover, to what extent should the distribution reflect other principles?
Having in the previous sections quoted different premises and considerations related to these questions, we shall now make an attempt at a more systematic and nuanced discussion. We shall all the time keep the principle of autonomy and Aristotle's three principles of constitutional justice in mind.
Let us first look at possible reasons for an `over-representation' of small (and medium-sized) countries.
First, ceteris paribus, as power index analyses indicate, a large state, under proportional voting weights, is likely to end up more often in a winning coalition than a small state, and less often in a losing coalition. This means, where cultural identity interests, or other vital interests, are affected, that the citizens of small states will experience heteronomy more often than do citizens of large states. Over-representation can reduce although not eliminate this difference.
Second, under proportional voting weights, large states need less extra votes than small states to arrive at a blocking, or a winning, coalition. Consequently, ceteris paribus, a large state can more easily than a small state arrive at a blocking, or a winning, coalition (cf. Johnston 1995). Over-representation of the small countries can reduce although not eliminate this difference.
There are also some other reasons why a large state, ceteris paribus, can more easily than a small state arrive at a blocking, or a winning, coalition. A large state is likely to have more capacity than a small state for the diplomatic ground-work needed; and because of its higher exchange value it probably need not invest that much effort to be listened to by other states, in particular other large states. Over-representation of the small countries can reduce although not eliminate the difference in exchange value.
Suppose now that for these, or other, reasons a scheme of voting weights has been adopted where small states are over-represented. Suppose then that the large states for some reason, e.g. because of the prospect of an enlargement that will bring in a number of small countries, demand redistribution in their favour. What kinds of arguments may be submitted? How can they be interpreted? What counter-arguments may be submitted?
Let us start out with arguments of the economic and financial realities type.
Such an argument may be of the oligarchic type; it may be an expression of the finance principle of constitutional proportionality. This principle, however, if taken seriously, will benefit countries with an efficient economy as much as it will benefit large countries. Moreover, although the financial principle of constitutional proportionality can be acceptable from a normative point of view in specialised international organisation of a technical or economic character, it cannot in a broadly oriented organisation like the EU with its emphasis on democratic principles.
However, an argument of the economic and financial realities type need not be of the oligarchic type.
Many issues that come up for decision in the EU institutions, in particular the Council, have a financial aspect. They require an allocation of money that is provided through taxation of the member states, i.e., ultimately, through taxation of the citizens of the union. Now, if each citizen contributes the same amount of money to the union, there is no a priori reason why citizens of some states are to have more (indirect) formal influence on allocations than citizens of other states. This, however, is a consequence of over-representation of small countries if the small states are not subject to a much harder taxation than the large states.
It is not surprising that German citizens argue for a redistribution of voting weights on such a basis. The prospect of an enlargement that implies an increased number of countries that are not only small (or medium-sized) but also clearly less affluent, may dramatise the kind of considerations outlined. The net contributors may find that, due to under-representation, they risk to be outvoted by the net receivers. According to Geoffrey Garrett (1992, p. 546), Germany and France, together with the BeNeLux countries, did not accept such a prospect when considering the decision rules in the planned SEA; they therefore demanded qualified majority voting instead of simple majority, which was the first priority of the southern, less affluent member states.
We have so far commented upon financial rather than economic realities. It is a question whether the economic significance, or power, of a large state may justify an increased voting weight for other reasons than the one that the country's economic capacity determines its contributions to the finances of the Union. Several lines of thought are possible. One may run like this: The quality of the economy of a large state may mean more to the community than the quality of the economy of a small state. In so far as a first-class understanding of the economy of a large state is likely to be higher within that state than outside it, the voting weight of the large state must not be too little. Indeed, the force of good arguments should, ideally, not depend upon voting weights, but in the real world voting power may be necessary to make people listen and take good arguments seriously.
Let us finally have a look at the idea submitted by the spokesman of Great Britain. He asserted that the driving force behind the European Union remains in the large countries, and submitted this assertion as an argument in favour of a redistribution of voting weight in the favour of the large states. His assertion is debatable. First, the various large states have rather different records as driving forces in the European Union. Second, it is a question in what respects and to what degree large states have proved to be more of a driving force than smaller states. Nevertheless, there may be something to the assertion. Just as the quality of the economy of a large state may mean more to the community than the quality of the community of a small state, the contribution of a large state, e.g., to security may mean more than that of a small state. This does not mean that a large state should have more formal power per citizen than a small state. It may mean, however, that if the under-representation of the large states goes too far, there will be an imbalance between voting power and possible or active substantive contributions, and there may be a demoralisation among the citizens. Admitting this, a person from a small state may, on the other hand, point out that another imbalance will quite easily arise. The large states, if resourceful enough in voting weights, may find it convenient to agree on policies among themselves without listening carefully to the interests and arguments of the smaller nations.
Finally a point of a different kind: if the under-representation of large states goes too far dissolution may be encouraged in countries where there are strong forces of secession.
(8) Some more considerations: winning and blocking
coalitions
The above reflections are very tentative, and they can hardly serve as a set of premises from which a scheme of voting weights can be deduced. Such reflections, however, may be of help when trying to assess a given scheme, e.g. the distribution of voting weights proposed by the Dutch Presidency under the IGC. More is needed, however, in order to arrive at a well-founded proposal that may gain acceptance; in particular, there is a need for clear analyses of what different schemes imply, or do not imply. There seem to be widespread erroneous beliefs that must be corrected and vague ideas that should be made more precise. In his recent article, The Voting System in the Council of the European Union: The Balance Between Large and Small Countries (1998), which is both critical and constructive, Axel Moberg (1998) makes a significant contribution in this respect. Among other things he points out that small and large countries are exposed to the same loss in relative power at each enlargement; and he demonstrates that the belief that small countries, after another enlargement, could outvote countries with a majority of the population is erroneous.
In this article, Moberg asserts that it is impossible to find an exact answer to what the large countries were trying to get at the conference (Moberg 1998, p. 350). His main conclusion is that blocking power was their central concern:
What the large countries seem to have been particularly preoccupied with at this conference is rather keeping the present (or restoring past) possibilities for a few of them of blocking a decision. This is important, since decision making in the EU operates largely through blocking minorities. The discussions of weighting of votes during the latest enlargement negotiations and the Ioannina compromise were just about that. (Moberg 1998, p. 351)
Even if this is not the whole truth it seems to me that abstract analyses should be supplemented with more concrete analyses of the conditions for blocking minorities and winning coalitions. Below, an attempt has been made to bring out what kinds of coalitions may constitute minimal winning coalitions, or blocking minorities. The intention is, for each of the historical stages, to contrast the winning coalition that is most clearly dominated by big powers, with its opposite. Similarly, the intention is, for each stage, to contrast the blocking coalition that is most clearly dominated by big powers, with its opposite. (It may be that I have not been able to pick out the most extreme coalitions.) Hopefully, such pairs of coalitions can help us think more realistically about the possible roles of large, medium-sized and small member states in the different stages. More concrete analyses, dealing with concrete countries and coalitions and with different types of issues, would be useful.
Some abbreviations:
L: Large state
Ll: Large state a little smaller than L
MM: Large medium-sized state
M: Medium-sized state
Mm: Medium-sized a little smaller than M
S: Small state
Ss: Very small state
Vw: Voting weight
QM: Qualified majority
Blm: Blocking minority
EC-6 (1958): 3L + 2M + 1Ss
Vw: (4,2,1)
QM: 3L 12/17
2L + 2M
Blm: 1L + 1M 6/17
EC-9 (1973): 4L + 2M + 2S + 1Ss
Vw: (10,5,3,2)
QM: 4L + 1Ss 41/58
3L + 1M + 1S
Blm: 2L or (1L+2M) 18/58
2M + 2S + 1Ss
EC-10 (1981): 4L + 3M + 2S + 1Ss
Vw: (10,5,3,2)
QM: 4L + 1M 45/63
3 L + 3M
Blm: 2L or (1L+2M) 19/63
3M + 1S + 1Ss
EC-12 (1986): 4L + 1Ll + 4M + 2S + 1Ss
Vw: (10,8,5,3,2)
QM: 4L + 1Ll + 1M + 1Ss 54/76
2L + 1Ll + 4M + 2S
Blm: 3 L or (2L+1M) 23/76
3M + 2S + 1Ss
EU-15 (1995): 4L + 1Ll + 4M + 2Mm + 3S + 1Ss
Vw: (10,8,5,4,3,2)
QM: 4L + 1Ll + 3M (+ 1Ss) 62-65/87
2L + 1Ll + 4M + 2Mm + 3S
Blm: 2L + 1Ll 26-23/87
2M + 2Mm + 2S (+ 1Ss)
EU-25 (?): 4L + 2Ll + 1MM+6M + 4Mm + 6S + 2Ss
Vw: (10,8,6,5,4,3,2)
QM: 4L + 2Ll + 7M + 1S + 1Ss 92/130
1L + 2Ll + 7M + 4Mm + 6S + 2Ss
Blm: 4L 39/130
2M + 2Mm + 6S + 2Ss
Conclusion
The proposal submitted by the Dutch presidency under the IGC contained both a double majority proposal, in two variants, and the single majority proposal that we have commented upon. I have limited my discussion to the single majority alternative. Now, some of the points made may be equally relevant in a discussion of a double majority system. I shall not, however, go into that possibility here. Let me just conclude by some remarks upon the single majority proposal submitted by the Dutch presidency.
This is a moderate proposal, so it ought to have some chance to win support in small member countries. At the same time, as I have shown, it can be linked to the set-up of EU-6, the original balance, in a way that might appeal to the large member states. As demonstrated by Moberg, a simple mathematical formula making the voting weight proportional to the square root of the population size - can cover the Dutch proposal, or a slight modification of it, fairly well (Moberg 1998, p. 563; cf. Philippe C. Schmitter & Jos� I. Torreblanca 1997). This adds to its attraction.
Notes
Footnotes
* Paper prepared for presentation at the ECPR Joint Sessions of Workshops at Mannheim 26-31 March, 1999
** ARENA and Department of Political Science, University of Oslo, P.O. Box 10967 Blindern, 0317 Oslo, Norway, e-mail: knut.midgaard@stv.uio.no
[1] I thank Madeleine O. Hosli, Aanund Hylland, Hannu Nurmi, Ulf Sverdrup and Martin S�ter for valuable help and advice at various stages, and I thank Ole-J�rgen Kvarsvik for stimulating and useful exchanges. I have moreover benefited from having an earlier version of the paper discussed at a workshop that was organised by the Swedish Political Science departments' network for research on Europe in Uppsala 24-26 August, 1998, and similarly at the ARENA annual conference, entitled The European Union and Small Nation States, in Oslo 5-6 November1998.
[2] Information from Ton de Bruijn in the Dutch Ministry of Foreign Affairs.
[3] This result was due to the author's co-operation with Aanund Hylland
[Date of publication in the ARENA
Working Paper series: 15.10.1999]